my experiments with technology

Error installing pytorch - OSError: [Errno 2] No such file or directory

2025-08-08T09:55:00.002-04:00

This error message: OSError: [Errno 2] No such file or directory is a common issue during software installations on Windows, especially with packages that have very long file paths, like PyTorch. The path you provided is extremely long, and Windows has a default limit on path length.

Here's a breakdown of the problem and the most common solutions:

The Problem: Long File Paths

The error isn't about PyTorch itself, but about a limitation in the Windows operating system. The default maximum path length on older versions of Windows is 260 characters. When the installation process tries to create a file at a path that exceeds this limit, it fails and throws this error. The path in your error message clearly exceeds this limit.

Solutions

Enable Long Path Support in Windows (Recommended): This is the most effective and permanent solution for this problem. It allows the operating system to handle file paths longer than 260 characters.
- For Windows 10/11 Home Edition:
  - Open the Registry Editor by pressing Win + R, typing regedit, and hitting Enter.
  - Navigate to: HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\FileSystem
  - Find the entry named LongPathsEnabled.
  - Double-click it and change the Value data from 0 to 1.
  - Restart your computer for the changes to take effect.
- For Windows 10/11 Pro/Enterprise Edition:
  - Open the Group Policy Editor by pressing Win + R, typing gpedit.msc, and hitting Enter.
  - Navigate to: Local Computer Policy > Computer Configuration > Administrative Templates > System > Filesystem
  - Double-click on "Enable Win32 long paths".
  - Select "Enabled" and click "OK".
  - Restart your computer.
Install to a Shorter Path: If you're unable to change the registry or group policy, you can try to install Python and its packages to a directory with a much shorter path. For example, instead of the default location in your user's AppData folder, you could install PyTorch and Python to C:\Python311 or a similar short path. You can use the --target flag with pip to specify a different installation directory.
Use a Virtual Environment: Using a virtual environment is a good practice for Python development in general. It can also sometimes help by creating a cleaner, and potentially shorter, installation path. However, if your virtual environment is still created in a directory with a very long path, you might still encounter the same error.

PyTorch on AMD GPUs

Since you mentioned you have an AMD-based laptop, it's also important to ensure you're installing the correct version of PyTorch that supports your GPU.

PyTorch for AMD GPUs on Windows and Linux relies on the ROCm (Radeon Open Compute) platform.
The official PyTorch website's "Get Started" page provides specific commands for installing PyTorch with ROCm support. You will need to select the appropriate options for your operating system and ROCm version.
On Windows, a common and recommended way to get GPU acceleration with AMD cards is to use the torch-directml package, which leverages DirectML. This is often easier to set up than ROCm on Windows. You can install it with the command pip install torch-directml.

Summary

The OSError: [Errno 2] No such file or directory is almost certainly due to the path length limit on your Windows system. The best solution is to enable long path support in Windows. After that, you can proceed with the standard PyTorch installation command. If you intend to use your AMD GPU for acceleration, make sure you're following the instructions for either the ROCm or DirectML backend, depending on your setup.

install google cloud api tools on ubuntu / debian

2023-10-01T00:13:00.001-04:00

What you need before Starting

Before the installation of the gcloud CLI, please confirm that the following conditions are met:

The OS release has not reached the end-of-life.
The apt-transport-http package is installed. If not, use the following command to install it:
```
sudo apt-get install apt-transport-https ca-certificates gnupg
```

Installation steps

Perform the following steps to install the Google Cloud CLI:

Step 1: Add `gcloud` CLI distribution

Use the following command to add the gcloud CLI distribution URI as a packet source:

Mounting Blob containers from Azure using blobfuse2

2023-04-03T22:45:00.001-04:00

Quite a few tutorials exist around the internet helping with tips on how to achieve this seemingly simple task. However, most of them choose the happy path and describe the simpler use cases.

Let me showcase some of the less visible tips -

1. if you are mounting into an already existing (non empty) directory, the mount call will fail
2. if your config file is picked up from baseconfig.yaml (from github page of blobfuse2), you need to take note of a few things -

a. the sample config file is not setup to allow other users to see the mounts / contents. You should change the entry "allow_other" to true from the example false. This change wont work unless you edit the /etc/fuse.conf file and uncomment the line #user_allow_other

b. the sample config file also indicates the storage type to be block, which works fine for flat blob storage situations. However, if you are using ADLS gen 2 for your storage account, you should indicate "adls"

3. Even though the config file has a container name entry, blobfuse2 allows you to override the container that you want to mount in the command line. for example -

blobfuse2 mount ${targetMountDirectory} --container-name ${containerName} --config-file=${configFilePath}

Apache Spark Interview Questions

2017-03-23T09:09:00.000-04:00

1. What is Spark?

Apache Spark is a fast, easy-to-use and flexible data processing framework. It has an advanced execution engine supporting cyclic data flow and in-memory computing. Spark can run on Hadoop, standalone or in the cloud and is capable of accessing diverse data sources including HDFS, HBase, Cassandra and others.

2. Key features of Spark.

Allows Integration with Hadoop and files included in HDFS.
Spark has an interactive language shell as it has an independent Scala (the language in which Spark is written) interpreter.
Spark consists of RDD’s (Resilient Distributed Datasets), which can be cached across computing nodes in a cluster.
Spark supports multiple analytic tools that are used for interactive query analysis , real-time analysis and graph

3. What is an RDD?

RDD is the acronym for Resilient Distribution Datasets – a fault-tolerant collection of operational elements that run parallel. The partitioned data in RDD is immutable and distributed. There are primarily two types of RDD:

Parallelized Collections : The existing RDD’s running parallel with one another.
Hadoop datasets : perform function on each file record in HDFS or other storage system

4. What are Partitions?

As the name suggests, partition is a smaller and logical division of data similar to ‘split’ in MapReduce. Partitioning is the process to derive logical units of data to speed up the processing process. Everything in Spark is a partitioned RDD.

5. What kind of operations RDD support?

Transformations.
Actions

6. What do you understand by Transformations in Spark?

Transformations are functions applied on RDD, resulting into another RDD. It does not execute until an action occurs. map() and filer() are examples of transformations, where the former applies the function passed to it on each element of RDD and results into another RDD. The filter() creates a new RDD by selecting elements form current RDD that pass function argument.

7. What are Actions in Spark ?

An action helps in bringing back the data from RDD to the local machine. An action’s execution is the result of all previously created transformations. reduce() is an action that implements the function passed again and again until one value if left. take() action takes all the values from RDD to local node.

8. Define functions of SparkCore?

Serving as the base engine, SparkCore performs various important functions like memory management, monitoring jobs, fault-tolerance, job scheduling and interaction with storage systems.

9. What is RDD Lineage?

Spark does not support data replication in the memory and thus, if any data is lost, it is rebuild using RDD lineage. RDD lineage is a process that reconstructs lost data partitions. The best is that RDD always remembers how to build from other datasets.

10. What is Spark Driver?

Spark Driver is the program that runs on the master node of the machine and declares transformations and actions on data RDDs. In simple terms, driver in Spark creates SparkContext, connected to a given Spark Master.
The driver also delivers the RDD graphs to Master, where the standalone cluster manager runs.

11. What is Hive on Spark?

Hive contains significant support for Apache Spark, wherein Hive execution is configured to Spark:

hive> set spark.home=/location/to/sparkHome;hive> set hive.execution.engine=spark;

Hive on Spark supports Spark on yarn mode by default.

12. Name commonly-used Spark Ecosystems.

Spark SQL (Shark)- for developers.
Spark Streaming for processing live data streams.
GraphX for generating and computing graphs.
MLlib (Machine Learning Algorithms).
SparkR to promote R Programming in Spark engine.

13. What is Spark Streaming ?

Spark supports stream processing – an extension to the Spark API , allowing stream processing of live data streams. The data from different sources like Flume, HDFS is streamed and finally processed to file systems, live dashboards and databases. It is similar to batch processing as the input data is divided into streams like batches.

14. What is Spark SQL?

SQL Spark, better known as Shark is a novel module introduced in Spark to work with structured data and perform structured data processing. Through this module, Spark executes relational SQL queries on the data. The core of the component supports an altogether different RDD called SchemaRDD, composed of rows objects and schema objects defining data type of each column in the row. It is similar to a table in relational database.

15. What is a Parquet file?

Parquet is a columnar format file supported by many other data processing systems. Spark SQL performs both read and write operations with Parquet file and consider it be one of the best big data analytics format so far.

16. What file systems Spark support?

Hadoop Distributed File System (HDFS). Learn more about HDFS in these Top Interview questions.
Local File system.
S3

17. What is Yarn?

Similar to Hadoop, Yarn is one of the key features in Spark, providing a central and resource management platform to deliver scalable operations across the cluster . Running Spark on Yarn necessitates a binary distribution of Spar as built on Yarn support.

18. List the functions of Spark SQL.?

Spark SQL is capable of:

Loading data from a variety of structured sources.
Querying data using SQL statements, both inside a Spark program and from external tools that connect to Spark SQL through standard database connectors (JDBC/ODBC). For instance, using business intelligence tools like Tableau. Get to know more about Tableau in this Tableau Tutorial.
Providing rich integration between SQL and regular Python/Java/Scala code, including the ability to join RDDs and SQL tables, expose custom functions in SQL, and more.

19. What are benefits of Spark over MapReduce?

Due to the availability of in-memory processing, Spark implements the processing around 10-100x faster than Hadoop MapReduce. MapReduce makes use of persistence storage for any of the data processing tasks.
Unlike Hadoop, Spark provides in-built libraries to perform multiple tasks form the same core like batch processing, Steaming, Machine learning, Interactive SQL queries. However, Hadoop only supports batch processing.
Hadoop is highly disk-dependent whereas Spark promotes caching and in-memory data storage.
Spark is capable of performing computations multiple times on the same dataset. This is called iterative computation while there is no iterative computing implemented by Hadoop.

Read more in this blog about the comparison of Spark and MapReduce.

20. What is Spark Executor?
When SparkContext connect to a cluster manager, it acquires an Executor on nodes in the cluster. Executors are Spark processes that run computations and store the data on the worker node. The final tasks by SparkContext are transferred to executors for their execution.

21. Name types of Cluster Managers in Spark.

The Spark framework supports three major types of Cluster Managers:

Standalone : a basic manager to set up a cluster.
Apache Mesos : generalized/commonly-used cluster manager, also runs Hadoop MapReduce and other applications.
Yarn : responsible for resource management in Hadoop

22. What do you understand by worker node?

Worker node refers to any node that can run the application code in a cluster.

23. Do you need to install Spark on all nodes of Yarn cluster while running Spark on Yarn?

No because Spark runs on top of Yarn.

Handling data with changing schema on Hadoop

2017-01-29T10:01:00.000-05:00

User data often is unpredictable, even when we can predict a change coming our way we need to prepare for that. Make changes in our environment to accept the incoming change, accommodate and absorb.

With that being a fact of life, our design should allow for those changes to happen. With the structured data warehouses gone (well, mostly), Hadoop based data stores are the norm these days, and would be default in future not so distant. By definition, Hadoop allows all kinds of data to be stored on the cluster and happily provides various tools to process such data.

However, with different types of data out there on cluster, any given tool would need to "know" what data that is? How to read it, interpret it and process it. That would have been relatively straightforward with the data schema being static, as in, a file comes in from source A, with a certain format, and that’s about it. You can define programs/structures to work with it.

What if the source system decides to alter the data set schema, add a few columns, and change some column ordering, perhaps data type too, what then? Would our program reading that dataset on Hadoop cluster still work as it? No.. It won’t .. Because it would also require change to reflect the same changes sent from source.

Assume that you did those changes too. So, now our program is version 2, able to handle the dataset sent by source (v2 as well). What about the older data lying around on Hadoop cluster (v1, with old schema)? How is that dataset going to used/usable? What if that data was used by hive tables, would you have two (or more) set of tables?

What if that data was exposed to some analytical programs, through R or similar? I am sure we won’t want to rewrite/test the whole solutions just because source system decided to change something?

And this scenario is not purely hypothetical, it happens around us, more often than not, just that we deal with it in our own patchwork styles, or solutions varying on a case to case basis.

Here I propose a somewhat generic solution, one that might work for all, well, nearly all use cases.

Avro is a data serialization system, wherein the schema is built into the file itself. As a result, programs reading the data-file doesn’t have to bother with interpreting the data file through external schema.

Avro was designed by Doug Cutting for similar reasons, as a tool that is language and schema independent and therefore unrelated tools/technologies can deal with data files.

That being said, is not a silver bullet in itself, but it handles the schema change fairly nicely.

Basically, you store the data on hdfs in Avro formatted files. For that, Avro schema needs to be created, which is fairly straightforward json structured file, indicating the record structure. Take a look at an example here -

{"namespace": "myExample.avro", "type": "record", "name": "User", "fields": [ {"name": "name", "type": "string"}, {"name": "age", "type": ["int", "null"]}, {"name": "address", "type": ["string", "null"]} ] }

create external table ex_users (name String , age int , address String) stored as AVRO location '/apps/data/hive/ex_users.tbl';

When files are saved as Avro formatted files, information from the schema is built into them, in plain text, and is parse-able for programs like hive and others to interpret the rest of the file.

For this to work, the tool has to rely on Avro. Look at this hive table definition -

With the clause "Stored as Avro" we are basically telling hive to match the columns defined in hive with columns in Avro fi

les in that hdfs location and use whatever columns map to hive definition.

Avro is compressible, split-able and stores data in binary format, providing some additional features.

When this Avro formatted file is exposed to programs like hive, they don’t have to worry about changes in schema, since they rely on the schema information in the file header. If the hive / R schema asks for 5 columns, but the file only has 3 of them (with names as matching criteria and not order/position of columns), those matching columns are project to the consuming tool, and the rest is ignored.

Similarly, if the tool asks for data that is not present in a certain data file (remember our example v2 file with additional column as against v1), the column that is not present is reported as null.

Now, look at the following data file (new avro schema - after source system decided to change)

{"namespace": "myExample.avro", "type": "record", "name": "User", "fields": [ {"name": "name", "type": "string"}, {"name": "age", "type": ["int", "null"]}, {"name": "salary", "type": ["float", "null"]}, {"name": "address", "type": ["string", "null"]} ]}

And the corresponding hive table definition -

create external table ex_users (name String , age int , address String, salary float)stored as AVROlocation '/apps/data/hive/ex_users.tbl';

For this table definition, its not a problem to query and report a data file in old schema (v1) wherein the column salary is not present. it would simply know from the avro schema definition in that header that the file doenst have that column, and it would gladly report that as null.

On the other hand, in the new structure has a different order of columns, which is also fine for hive/avro combination since the schema definition helps match up the columns and report the right content in right columns.

Similar analogies can be derived for other data consumers who need to interpret the schemas for data files on hdfs.

In comparison to some other storage formats, we do lose some points on the performance scale, but for the use case, where schema is changing and we want to be able to handle the changes with little or no effort, this is a better fit than many others.

Hadoop - Small Files vs Big Files

2016-12-20T10:26:00.000-05:00

Credits- https://blogs.msdn.microsoft.com/cindygross/2015/05/04/hadoop-likes-big-files/

One of the frequently overlooked yet essential best practices for Hadoop is to prefer fewer, bigger files over more, smaller files. How small is too small and how many is too many? How do you stitch together all those small Internet of Things files into files "big enough" for Hadoop to process efficiently?

The Problem

One performance best practice for Hadoop is to have fewer large files as opposed to large numbers of small files. A related best practice is to not partition “too much”. Part of the reason for not over-partitioning is that it generally leads to larger numbers of smaller files.

Too small is smaller than HDFS block size (chunk size), or realistically small is something less than several times larger than chunk size. A very, very rough rule of thumb is files should be at least 1GB each and no more than maybe around 10,000-ish files per table. These numbers, especially the maximum total number of files per table, vary depending on many factors. However, it gives you a reference point. The 1GB is based on multiples of the chunk size while the 2nd is honestly a bit of a guess based on a typical small cluster.

Why Is It Important?

One reason for this recommendation is that Hadoop’s name node service keep track of all the files and where the internal chunks of the individual files are. The more files it has to track the more memory it needs on the head node and the longer it takes to build a job execution plan. The number and size of files also affects how memory is used on each node.

Let’s say your chunk size is 256MB. That’s the maximum size of each piece of the file that Hadoop will store per node. So if you have 10 nodes and a single 1GB file it would be split into 4 chunks of 256MB each and stored on 4 of those nodes (I’m ignoring the replication factor for this discussion). If you have 1000 files that are 1MB each (still a total data size of ~1GB) then every one of those files is a separate chunk and 1000 chunks are spread across those 10 nodes. NOTE: In Azure and WASB this happens somewhat differently behind the scenes – the data isn’t physically chunked up when initially stored but rather chunked up at the time a job runs.

With the single 1GB file the name node has 5 things to keep track of – the logical file plus the 4 physical chunks and their associated physical locations. With 1000 smaller files the name node has to track the logical file plus 1000 physical chunks and their physical locations. That uses more memory and results in more work when the head node service uses the file location information to build out the plan for how it will split out any Hadoop job into tasks across the many nodes. When we’re talking about systems that often have TBs or PBs of data the difference between small and large files can add up quickly.

The other problem comes at the time that the data is read by a Hadoop job. When the job runs on each node it loads the files the task tracker identified for it to work with into memory on that local node (in WASB the chunking is done at this point). When there are more files to be read for the same amount of data it results in more work and slower execution time for each task within each job. Sometimes you will see hard errors when operating system limits are hit related to the number of open files. There is also more internal work involved in reading the larger number of files and combining the data.

Stitching

There are several options for stitching files together.

Combine the files as they land using the code that moves the files. This is the most performant and efficient method in most cases.
INSERT into new Hive tables (directories) which creates larger files under the covers. The output file size can be controlled with settings like hive.merge.smallfiles.avgsize and hive.merge.size.per.task.
Use a combiner in Pig to load the many small files into bigger splits.
Use the HDFS FileSystem Concat APIhttp://hadoop.apache.org/docs/current/api/org/apache/hadoop/fs/FileSystem.html#concat.
Write custom stitching code and make it a JAR.
Enable the Hadoop Archive (HAR). This is not very efficient for this scenario but I am including it for completeness.

There are several writeups out there that address the details of each of these methods so I won’t repeat them.

Merging small files on HDInsight http://blogs.msdn.com/b/mostlytrue/archive/2014/04/10/merging-small-files-on-hdinsight.aspx which uses a Java MapReduce JAR https://github.com/mooso/smallfilesmerge.
Quick Tip for Compressing Many Small Text Files within HDFS via Pighttp://dennyglee.com/2014/01/06/quick-tip-for-compressing-many-small-text-files-within-hdfs-via-pig/.
FileCrush https://github.com/edwardcapriolo/filecrush.
HDFS FileSystem Concat API
CombineFileInputFormat (splits)
- This may not work with really large numbers of files and has to be used EVERY time a job is run.
- http://www.ibm.com/developerworks/library/bd-hadoopcombine/index.html
- Process Small Files on Hadoop Using CombineFileInputFormat (1)http://www.idryman.org/blog/2013/09/22/process-small-files-on-hadoop-using-combinefileinputformat-1/
Dealing with Hadoop’s small files problem http://snowplowanalytics.com/blog/2013/05/30/dealing-with-hadoops-small-files-problem/ “aggregating with the small files first reduced total processing time from 2 hours 57 minutes to just 9 minutes – of which 3 minutes was the aggregation, and 4 minutes was running our actual Enrichment process. That’s a speedup of 1,867%.”
The Small Files problem in Hadoop http://piglog4j.blogspot.com/2013/06/the-small-files-problem-in-hadoop.html
Hadoop Archive: File Compaction for HDFS https://developer.yahoo.com/blogs/hadoop/hadoop-archive-file-compaction-hdfs-461.html
The Small Files Problem http://blog.cloudera.com/blog/2009/02/the-small-files-problem/ “Reading through files in a HAR is no more efficient than reading through files in HDFS, and in fact may be slower since each HAR file access requires two index file reads as well as the data file read (see diagram). And although HAR files can be used as input to MapReduce, there is no special magic that allows maps to operate over all the files in the HAR co-resident on a HDFS block.”

The key here is to work with fewer, larger files as much as possible in Hadoop. The exact steps to get there will vary depending on your specific scenario.

Eclipse - installing Scala plugin manually?

2016-11-15T16:33:00.000-05:00

I have been playing around with Scala for some time, and was always using the Scala IDE (www.scala-ide.org) which is based on a relatively older version of Eclipse (Luna).

I recently discovered this, wherein you could install the scala plug-in on a regular Eclipse installation.

Just add the following url as a new update site in your local eclipse installation and you'd be able to install the scala plugin just like that -

http://download.scala-ide.org/sdk/lithium/e44/scala211/stable/site

Links to free big-data-sets

2016-07-23T20:42:00.000-04:00

Many people who are starting their journey with big data and analytics find it hard to get their hands on the right kind of data to play or experiment with.

Most of the time, people have enthusiasm, they are learning the skill too, but they just don't have the right kind of dataset to apply their newly acquired skills.

Democratising data has been at the forefront of discussions for many data pioneers. Through their efforts and with some re-alignment of technology priorities, some government bodies have opened up their datasets to the public.

As a result, here is a set of links (reproduced) to some of the free sources.

Data.gov http://data.gov The US Government pledged last year to make all government data available freely online. This site is the first stage and acts as a portal to all sorts of amazing information on everything from climate to crime.
US Census Bureau http://www.census.gov/data.html A wealth of information on the lives of US citizens covering population data, geographic data and education.
Socrata is another interesting place to explore government-related data, with some visualisation tools built-in.
European Union Open Data Portal http://open-data.europa.eu/en/data/ As the above, but based on data from European Union institutions.
Data.gov.uk http://data.gov.uk/ Data from the UK Government, including the British National Bibliography – metadata on all UK books and publications since 1950.
Canada Open Data is a pilot project with many government and geospatial datasets.
Datacatalogs.org offers open government data from US, EU, Canada, CKAN, and more.
The CIA World Factbook https://www.cia.gov/library/publications/the-world-factbook/Information on history, population, economy, government, infrastructure and military of 267 countries.
Healthdata.gov https://www.healthdata.gov/ 125 years of US healthcare data including claim-level Medicare data, epidemiology and population statistics.
NHS Health and Social Care Information Centre http://www.hscic.gov.uk/home Health data sets from the UK National Health Service.
UNICEF offers statistics on the situation of women and children worldwide.
World Health Organization offers world hunger, health, and disease statistics.
Amazon Web Services public datasets http://aws.amazon.com/datasets Huge resource of public data, including the 1000 Genome Project, an attempt to build the most comprehensive database of human genetic information and NASA ’s database of satellite imagery of Earth.
Facebook FB +0.32% Graph https://developers.facebook.com/docs/graph-api Although much of the information on users’ Facebook profile is private, a lot isn’t – Facebook provide the Graph API as a way of querying the huge amount of information that its users are happy to share with the world (or can’t hide because they haven’t worked out how the privacy settings work).
Face.com: A fascinating tool for facial recognition data.
UCLA makes some of the data from its courses public.
Data Market is a place to check out data related to economics, healthcare, food and agriculture, and the automotive industry.
Google Public data explorer includes data from world development indicators, OECD, and human development indicators, mostly related to economics data and the world.
Junar is a data scraping service that also includes data feeds.
Buzzdata is a social data sharing service that allows you to upload your own data and connect with others who are uploading their data.
Gapminder http://www.gapminder.org/data/ Compilation of data from sources including the World Health Organization and World Bank covering economic, medical and social statistics from around the world.
Google GOOGL +0.66% Trends http://www.google.com/trends/explore Statistics on search volume (as a proportion of total search) for any given term, since 2004.
Google Finance https://www.google.com/finance 40 years’ worth of stock market data, updated in real time.
Google Books Ngrams http://storage.googleapis.com/books/ngrams/books/datasetsv2.htmlSearch and analyze the full text of any of the millions of books digitised as part of the Google Books project.
National Climatic Data Center http://www.ncdc.noaa.gov/data-access/quick-links#loc-clim Huge collection of environmental, meteorological and climate data sets from the US National Climatic Data Center. The world’s largest archive of weather data.
DBPedia http://wiki.dbpedia.org Wikipedia is comprised of millions of pieces of data, structured and unstructured on every subject under the sun. DBPedia is an ambitious project to catalogue and create a public, freely distributable database allowing anyone to analyze this data.
New York Times http://developer.nytimes.com/docs Searchable, indexed archive of news articles going back to 1851.
Freebase http://www.freebase.com/ A community-compiled database of structured data about people, places and things, with over 45 million entries.
Million Song Data Set http://aws.amazon.com/datasets/6468931156960467 Metadata on over a million songs and pieces of music. Part of Amazon Web Services.
UCI Machine Learning Repository is a dataset specifically pre-processed for machine learning.
Financial Data Finder at OSU offers a large catalog of financial data sets.
Pew Research Center offers its raw data from its fascinating research into American life.
The BROAD Institute offers a number of cancer-related datasets.

Credit to Forbes article at

http://www.forbes.com/sites/bernardmarr/2016/02/12/big-data-35-brilliant-and-free-data-sources-for-2016/#5b2a54cf6796

Teradata Data type abbreviation - described

2015-06-19T10:59:00.002-04:00

Teradata data types (as reported in DBC.Columns.ColumnType can be cryptic and not always easy to remember. Here's a ready reckoner -

Abbreviation	Equivalent English :)
A1	ARRAY
AN	MULTI-DIMENSIONAL ARRAY
AT	TIME
BF	BYTE
BO	BLOB
BV	VARBYTE
CF	CHARACTER
CO	CLOB
CV	VARCHAR
D	DECIMAL
DA	DATE
DH	INTERVAL DAY TO HOUR
DM	INTERVAL DAY TO MINUTE
DS	INTERVAL DAY TO SECOND
DY	INTERVAL DAY
F	FLOAT
HM	INTERVAL HOUR TO MINUTE
HS	INTERVAL HOUR TO SECOND
HR	INTERVAL HOUR
I	INTEGER
I1	BYTEINT
I2	SMALLINT
I8	BIGINT
JN	JSON
MI	INTERVAL MINUTE
MO	INTERVAL MONTH
MS	INTERVAL MINUTE TO SECOND
N	NUMBER
PD	PERIOD(DATE)
PM	PERIOD(TIMESTAMP WITH TIME ZONE)
PS	PERIOD(TIMESTAMP)
PT	PERIOD(TIME)
PZ	PERIOD(TIME WITH TIME ZONE)
SC	INTERVAL SECOND
SZ	TIMESTAMP WITH TIME ZONE
TS	TIMESTAMP
TZ	TIME WITH TIME ZONE
UT	UDT Type
XM	XML
YM	INTERVAL YEAR TO MONTH
YR	INTERVAL YEAR
=++	TD_ANYTYPE

Hadoop Meetup on the sidelines of Strata Hadoop Conference - Part 2

2015-05-06T16:52:00.001-04:00

Read part 1 of this here

Day 2 of the meetup was equally exciting, if not better. Lined up were talks from Qubit and Google, William Hill (a surprise for me - more later on that) and then PostCodeAnywhere, all very exciting from the synopsis.

Google & Qubit showcased basically a stream processing engine, with pluggable components, many of them can be written in different technologies and programming languages.

Of course Google Cloud Data flow is much more than just a stream processing engine, however, for real time data ingestion perspective, that feature is pretty significant.

A completely managed system, it woks on the publish-subscribe (pub-sub) model. As Reza put it, “pub-sub is not just data delivery mechanism, its used as a glue to hold the complete system together”. Pluggable components is another differentiator for Google’s offering, in today’s demo they showcased bigtable as one of the consumers at the end.

From my own knowledge of stream processing, which is not significant in anyway, i could relate to many similarities with IBM’s info sphere streams and some with apache kafka. However, a question around comparisons with these sites remained unanswered from Google (though in very good spirit, in a chat with the speaker Reza later on, it came out as more of a philosophical question avoidance than anything else).

The william hill talk (by Peter Morgan, their head of engineering), was a genuine surprise, at least for me. Perhaps due to my ignorance, due to which i didn't realize, their systems are far more sophisticated and load bearing than I would have imagined. As an example, they process 160TB of data through their systems on a daily basis.

Including many complexities managed through their system are their main components, the betting engine, the settlement engine among others.

William Hill supports an open API as well, enabling app developers to pick up data elements and innovate. However, for obvious reasons, very limited data is thrown open in the public domain. Would that be a deterrent for app developers ? not having enough data ? For example, if i would want to report in an app, who’s betting on a certain game, cross referenced with geo location data .. I cant do that, since William hill doesn't publish demographic data. I personally feel alright with it, there are possibilities that many of those data elements can be used in ways to influence the betting system itself, becoming counter-productive.

I would imagine their IT systems to be one of the top notch systems around the place, to be able to manage such data volumes, with such speeds and accuracy. Commendable job. I would probably write exclusively on their architecture once i get my hands on the presentation slides (couple of days may be).

The talk from PostCodeAnywhere was more educative to me, personally. Got to understand a bit about Markov Models, chains and how they can be used for machine Learning. Very interesting stuff there too.

Apache Spark is being seen more and more as the tool to be perform analytics on the fly, specially on large volumes of data. It would be very interesting to see how R and python analytical capabilities compare with what spark offers.

Speaking to another attendee today, it came out the people prefer to use R more and more for massaging and cleansing purposes, however, its not seen as fit for heavy lifting required for performing real analytic and/or predictive pieces. For these areas, people still prefer to use Python.

IBM’s bigR is a possible contender for the job, where they talk about having optimised R for a hadoop cluster and have enabled it to work on top of hdfs. However, bigR is not open source and that could be its biggest challenge in adoption.

Hadoop Meetup on the sidelines of Strata Hadoop Conference - Part 1

2015-05-06T03:48:00.001-04:00

Not being able to make it up to the Main conference (Strata Hadoop London 2015), the evening meet-ups were the consolation pieces of getting in touch as much as possible.

In my view, these conferences/events often help one to get to know about the recent developments in the space, mostly showcasing whats being done with a given technology, whats coming up (future developments, innovations) and people's experiences with the technology, both good (the famous savings use cases) and bad (challenges faced in achieving production readiness, if ever).

Last evening, on day 1 of the conference, I ended up attending one of the meet-ups. It was particularly useful for me, for couple of reasons.

There was a talk on the new execution engine for hive, i.e. hive running on spark. Always keen on internal workings of a complex piece of software (or hardware for that matter), i was very happy to be able to listen in directly from the person responsible for much of development on hive. I have an audio recording of the whole talk, though I am hopeful that the conference organizers shall put up the video on their website anyway.

When Phill talked about his experiences on getting hadoop on its feet and how they orchestrated hadoop as a PaaS within BT was something insightful, (they seem to call it HaaS there). it showed two things to me - architects' always have to "find the funding" for innovations and new tech to be brought in to the organization :) Also, Security on hadoop is "doable", as his use case proved to be. There are reliable tools and solutions which can help achieve Enterprise level security for a hadoop cluster.

Another interesting talk was Dato's. Dato is a machine learning/modelling tool, which claims to be fairly quick than many others, allows the data to be consumed in-place (like hadoop) and supports hdfs integration. I am sure to follow up on Dato with the organization. for me its one of the key problems of the future, where data is too much and the modelling algorithm has to be enabled to consume data for training sets in-place, since its just not practical to move tera/petabytes of data to where program is. IBM BigR is doing something similar as well.

Finally, another interesting talk was from Big Data Boards team. they talked about how they are building cluster hardware for hosting small hadoop clusters. Interesting proposition there, to have your own hadoop cluster running on a desk in a corner of your office. no need of going over to the likes of aws for hosting the cluster. They say that many universities etc are already using the clusters they made for real life experiments. very interesting space for me.

Data Sets

2015-04-04T03:49:00.000-04:00

Some of the publicly available datasets are listed here. This is a continuously evolving page, and therefore might not always be 100% up-to-date. For licensing information, please refer to each datasets’ own licensing page. I take no responsibility for the licensing/distribution of the datasets.

1. Amazon’s Ratings dataset

http://jmcauley.ucsd.edu/data/amazon/

2. imdb movies dataset

http://www.imdb.com/interfaces

3. London Transport Data - tfl

http://www.tfl.gov.uk/info-for/open-data-users/our-feeds?intcmp=3671#on-this-page-0

Weighted trust graph for authentication

2015-02-22T17:17:00.000-05:00

During the hackathon (discussed in earlier post), I met with Gary. He had come in to be play mentor, but since he couldn't devote enough time, ended up being a guest.

After engaging Neo4J in few meetups and understanding the database a bit, i was contemplating using Graph databases for authentication. It might have consequent applications in fraud analytics too, where graph databases are used already[1]

During the discussion on the idea, Gary suggested to mould the idea differently and possibly using Trust networks/graphs, wherein each node (entities i.e. people, organisations etc) are related to each other through weighted directed relationships. The weight of this relationship can be deduced in multiple ways, e.g. by periodic algorithms similar to search engine ranking algorithms or by asking people their trust level of others on a scale of 1 to x. x being a hypothetical standard scale that can be used as a yardstick across the network for determining level of trust.

While researching some more, I found that similar research has been done in this space [2], though applications are few to come by.

It was also pointed out in the discussion that banks don't really have a huge interest in preventing this crime. the view was, since banks already provision for certain amount in their balance sheets for these "potential" thefts, they don't really bother so much.

I believe that the financial institutions as a single unit need to attack these fraud crimes by joining hands and leveraging best of research and technology for minimising the crime. The technology exists for providing unto the moment information on these events, some more innovation and research is needed that can bring together the whole picture and look like a "solution"

[1]http://info.neotechnology.com/rs/neotechnology/images/Fraud%20Detection%20Using%20GraphDB%20-%202014.pdf?_ga=1.182367911.1656585956.1417700858
[2]http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.212.9978&rep=rep1&type=pdf

Hackathon - Fintechathon

2015-02-16T17:53:00.002-05:00

A hackathon is a hacking marathon wherein many people are invited to attack problems around a theme.

I recently attended a hackathon over the valentines weekend. Organised by StartupBootCamp Fintech, it was attended by about 100 people. Many ideas, many teams, some partners i.e. corporates with their own challenges.

I was initially team less, but then found some others who were in in my situation. We formed a team, around my favourite topic, data analytics. We had two business development guys, Adam and Oksana, two java programmers, Nelson and Nick, and a mobile app developer Vlad.

Hackathon teams are formed around ideas, wherein someone with an idea takes on the ownership, and the collects the team around it. Things are focussed from moment one, and the march forward is fairly disciplined and fast, thats why the name hackathon... keep hacking, for long, long days and nights.

We, had the other way round.. all of us were teamless and therefore put together as a team. We had no idea to start with. As a result of that, we spent better part of the friday evening and saturday zeroing on the problem to attack.

Finally we decided to go ahead with an data analytics piece. I won't chalk out the details here, but its something that the marketing guys always love and like. To know when their customers are off to a life event, and therefore could be offered some product.

By Sunday morning, we had lost two team members, one to a different idea and one to sleep. Vlad hadn't slept in 4 nights, so he kept sleeping much of Sunday.

As a result, we ended up a team without anyone who could do any UI design, and therefore only some backend API calls, some analytics pieces and nothing to show off.

The result was that, we couldn't show any working model in our pitch presentation and had to contend with a presentation only, which tried to describe our idea to the judges.

Of course we lost, but then it was a very nicely spent weekend, met some very nice people, made some contacts, and possibly a future for the idea.

Playing with R

2015-01-31T18:07:00.000-05:00

Since early 2013, when I got to know of possible online learning for R through www.coursera.com, I have been a fan of this language.

My teachers in college would probably not like to hear this, but I was never a great Statistics student, it was always a "passing" requirement during my studies. However, when I learned about R and did some experimental work with R, i found that stats is indeed a very interesting and powerful mathematical tool. And with R i found it fun too, which in my personal opinion is the root of an ongoing learning process. If learning is not fun, it stops.

For me, laying my hands on RStudio was the best thing that happened to me around the statistics learning curve. The tool makes your R learning curve so much more easy and possible, the college could never do that.

As an example, in college we could never realise real life examples and uses of regression, curve fitting etc. However, now using R, it all makes so much more sense and with lot of ease.

I can't put a finger on the clear reason, whether the years have added that capacity to understand the subject, or the very fact that today these kinds of tools exist which let you explore that area with so much straightforward ease, but the fact remains. It has become so much easy, relatively, to get your hands on the power of statistics and analytics with tools like.

Thats probably one reason, why i am getting more and more inclined towards data analytics. Its a very powerful and interesting area, with huge potential towards shaping our future.

Can we rollback the Truncate, Drop, and Delete command deleted data?

2014-05-26T07:44:00.001-04:00

Here we talk about the concept of truncate, drop and delete commands and whether they can be rollback, if so, how and when !!

Yesterday I was working on this and see if we use the truncate command in transaction, we can rollback the data. One more thing if we perform delete command outside the transaction, we can’t rollback the data.

Let’s try this…

Suppose I have a table dbo.Employee with three column- Id, Name, and Salary.

CREATE TABLE [dbo].Employee(
      Id [int] NULL,
      Name [varchar](50) NULL,
      Salary [Int] NULL
)

Insert the values into the db.Employee table

INSERT INTO dbo.Employee VALUES(1,'Employee1',10000)
INSERT INTO dbo.Employee VALUES (2,'Employee2',20000)
INSERT INTO dbo.Employee VALUES (3,'Employee3',30000)
INSERT INTO dbo.Employee VALUES (4,'Employee4',10000)
INSERT INTO dbo.Employee VALUES (5,'Employee5',30000)

Now I have a five rows in table

SELECT * FROM dbo.Employee

Now we perform the delete command in Transaction and then select command

BEGIN TRAN
DELETE FROM dbo.Employee WHERE ID = 2
ROLLBACK

SELECT * FROM dbo.Employee

Output:

But again I run the select command then I get the same result as previous select statement (i.e 5).
Now I perform the Truncate command in Transaction, Let see what happen.

BEGIN TRAN
TRUNCATE TABLE dbo.Employee
ROLLBACK

Now again I run the select command and got the same rows and same result, if same thing we do with Drop command that will also give same result.
It means If we use the Delete, Truncate, and Drop command with Transaction, we can rollback the data and if we perform all three command outside the Transaction we can’t rollback.

One more thing I have to add in this article, we can’t rollback the data but we can restore the data from the Transaction_log, if we use the delete command. We can’t restore the data if we use the truncate or drop command. How and Why read in next article.

In summary,
We can rollback the DDL and DML commands, if all these commands are in transaction. We can only restore the DML commands data.

Difference between Char, Varchar, and nVarchar data type in SQL Server

2014-05-22T04:53:00.001-04:00

In real world everyone has different need and requirement, why we use the same data type to store different types of value. What is the actual difference between Char, Varchar and nVarchar data type.

Char DataType

Char datatype which is used to store the fixed length, non-unicode characters data. Suppose if we declared char (50) it will allocates memory for 50 characters. Once we declare char (50) and if we enter fewer than the number of characters defined (i.e. 6), the remaining length will be space filled to the right. it means other 44 characters of memory will be wasted.

Now we look in the query, How will it effect.

Output of this Select Statement:

Use this data type when the column data is to be of fixed length, which tends to be the case

for customer IDs and bank account IDs.

Nchar DataType

Nchar type is exactly similar to char. But hold character in Unicode format rather than ANSI. The Unicode format has a larger character set range than ANSI. Unicode datatype take exact double memory space in sql server. So use the nchar when you want to store the Unicode value.

varchar DataType

Varchar datatype, which is used to store the alphanumeric values, just like char datatype. If both are similar than what is the difference between char and Varchar? What is the need to create the new datatype? Later in this article we will discuss this. The maximum size of a Varchar column is 8,000 characters. However, if you define the column with no size—that is, Varchar () — then the length will default to 1.

Difference between the char and Varchar datatype

Char belongs to fixed length characters datatype, Varchar belongs to variable length characters datatype. If you define a column to be 20 characters long, but If you enter fewer than the number of characters defined like 10 characters then it will consume only 10 characters memory. It will not consume the defined memory.

Now we look in the query:

Output of the select statement:

We can see in above image the output of the select statement, the ZipCodelength value are varying according to data in zipCode field.

So we should use the Char for the fixed length data field and Varchar for the variable length data field.

Nvarchar DataType

Nvarchar datatype same as varchar datatype but only difference nvarchar is used to store Unicode characters and it allows you to store multiple languages in database. nvarchar datatype will take twice as much space to store extended set of characters as required by other languages.

Difference between Varchar and Nvarchar DataType

Varchar(n)

Nvarchar(n)

1. Non-Unicode Variable Length character data type.

Example: Declare @FirstName As Varchar(20) = ‘Developer’

Select @FirstName

2. It takes 1 byte per character.

Example: Declare @FirstName As Varchar(20) = ‘Developer’

Select @FirstName AS FirstName,

DATALENGTH(@FirstName) As Length

Result:

FirstName Length

Developer 9

3. Can store maximum 8000 Non-Unicode characters.

4. If the value of n is not specified at the time of variable declaration or column definition then it is considered as 1.

Example: Declare @FirstName As Varchar =’Developer’

SELECT @firstName FirstName,

DATALENGTH(@firstName) Length

Result:

FirstName Length

D 1

5. If n is not specified in while using Cast/Convert functions, then it is considered as 30.

Example:

Declare @firstName Varchar(50)=
‘Unicode character sets hold up to 65,536’

Select CAST(@firstName As Varchar) FirstName,
DATALENGTH(CAST(@firstName As varchar)) Length

Result:
FirstName Length
Unicode character sets hold up 30

1. Unicode Variable Length character data type. Nvarchar can store both the Unicode and non-Unicode (i.e) Chinese, Japanese, Korean etc) character.

Example: Declare @FirstName As Nvarchar(20) = ‘Developer’

Select @FirstName

2. It takes exactly double bytes per character as compare to Varchar. It takes 2 bytes per Unicode/Non-Unicode character.

Example: Declare @FirstName As Nvarchar(20)= ‘Developer

Select @FirstName AS FirstName,

DATALENGTH(@FirstName) AS Length

Result:

FirstName Length

Developer 18

3. Can store maximum 4000 Unicode/Non-Unicode characters.

4. If the value of n is not specified at the time of variable declaration or column definition then it is considered as 1.

Example: Declare @FirstName As Nvarchar =’Developer’

SELECT @firstName FirstName,

DATALENGTH(@firstName) Length

Result:

FirstName Length

D 2

5. If n is not specified in while using Cast/Convert functions, then it is considered as 30.

Example:

       Declare @firstName Nvarchar(50)=
‘Unicode character sets hold up to 65,536’

Select CAST(@firstName As Nvarchar) FirstName,
DATALENGTH(CAST(@firstName As Nvarchar)) Length
Result:
FirstName                                 Length
Unicode character sets hold up   60

So if you are not using other languages then it’s better to use varchar datatype instead of nvarchar.

Welcoming Microsoft world onto my blog...

2014-05-20T08:31:00.000-04:00

For most of my IT life, I have been thoroughly disconnected with the Microsoft world of enterprise products. One of the reasons could be the fact that I was introduced into the IT through the corridors of Java & then Unix.. Subsequently more things came on the way, but I never got around to diving deep in the MS world.

Sometime back, I did try and explored SQL server 2008 R2 and associated ETL/Analytics pieces, but it didnt excite me as much.

Now, I have an opportunity to learn and share the MS wisdom through a dear friend Sachin Dubey. He's going to be putting up content on to this blog. He would be contributing content geared towards Microsoft products, e.g. SQL Server, MS-BI, Dot-Net etc. With significant experience working around MS BI space, Sachin is a self built knowledge repository. I hope we all learn with him.

Programmatic interaction with Facebook data

2014-02-08T18:30:00.000-05:00

The last time I interacted with Facebook using a program (2011), it was fairly straightforward, and perhaps it was too simple, it was abused. As a result, in my recent experiments, I come to find out that, they have made it so complicated (my feeling, no offence to Facebook here), that its quite hard, if not impossible to do genuine development work using Facebook API, let alone achieve something sinister.

I am a self-anointed tech junky, and therefore no one should be surprised that i took up this challenge couple of hours back to build a brand new routine to get data in/out from/to facebook using a popular and open source Data integration tool, called Pentaho

There were two triggers, one, I had not done some core debugging based research using something new. Second, a friend of mine is building this ecommerce site, www.freemall.in for which the marketing efforts would need a simpler interface to push details to the site, instead of some human going up there and doing all. So, this programming interface becomes the first step of the future plugin for his ecommerce venture.

As I already indicated, programming with facebook's current API is a relatively hard task, at least for couple of iterations, before you get the hang of it. So, I started with the tutorial at pentaho's site - http://www.pentahoevalcenter.com/data-integration/advanced-data-integration/facebook/ which turns out to be based on the old Facebook API and prompted me to look elsewhere for the right way of working.

Then comes the big thing - the oAuth authentication protocol employed by Facebook, get a token of this type, then of this type. go to this site, achieve this, copy this and take it there.. Its all so much that I had to go back to google and see if there is something simpler.. and yes, god bless google, there was someone like me who had gone to all the trouble and then of documenting it... God bless his soul for this -
http://stackoverflow.com/questions/12168452/long-lasting-fb-access-token-for-server-to-pull-fb-page-info

After this, I went ahead with the pentaho tutorial, which did work out nice and fine for me.

I was able to post couple of messages to my test page using the application that i built for this test case.

hope this collection of links at a single place helps others.

Proud - Cloudera Certified Developer for Apache Hadoop

2014-02-02T15:46:00.000-05:00

I have not been a believer of certifications, that should be clear from the fact that even though I have been working with Informatica since 2002, I never tried their certification till 2011. And, same about Oracle, whose exam I have not yet attempted/planned/thought over. I used to think that the knowledge level will prevail anyway, whether or not an authority stamps on it.

However, I believe, I am changing, to a certain extent so to say. It seems that I have come to accept the certifications' worth, and therefore, after the training provided by Cloudera, I picked up the opportunity and went through with the rigor of examination.

Fortunately, I came through. And, as much as I try not to showcase it, its a great feeling. Somehow, the knowledge is vindicated, that yes, this guy knows something about hadoop and you better listen to him, :) . Funny that one has to put a badge out there to be heard.

Well, all said and done, the certification is done, courtesy the employer, who sponsored the training and of course the examination coupon that came along. I would like to thank our trainer from cloudera, Amandeep Khurana for his depth and breadth around all things hadoop. As much as I knew about hadoop before going in the training, those 2.5 days added precious layers to my knowledge. Thanks Amandeep.

Removing ports from an existing mapping - Do's and Dont's

2014-01-30T03:58:00.000-05:00

Recently, a colleague called in for an issue that he was facing with an Informatica mapping. Let me recreate that situation here -

There is a mapping that get n ports from a source and loads m ports to a target. Standard stuff, nothing special or fishy around there. There comes a change request that says, such and such x number of columns are not required in the target, and since there would be a sure performance penalty for carrying through extra data (however small), the mapping should be changed to remove those ports from the pipeline.

What this gentleman did was to remove the connections for those ports from the source qualifier onwards. That saved him from changing and re-importing the source definition.

However, the mapping execution failed, complaining about a certain error situation. Thats when I got the call :)

whether I could discuss and fix it is something else, but what was the reason of the error - ?

The very fact that if you are bringing in certain ports in the source qualifier, you HAVE to take them forward. Thats a rule from informatica's side. Which means, if you are pulling up n ports from the source, you HAVE to expose all those n ports going away from the source qualifier. Or, putting it differently, in whichever way you create a set of ports in your source qualifier, all of them have to be consumed by some transformation object. Not a single one can be left unconnected, on the input side or the output side. Simple.

That was the reason, and the solution was fairly simple, to reduce his effort, either remove the ports from source qualifier as well, or just carry them forward to one transformation. and then drop them onwards.

Having said that, the very reason for which this whole change was initiated, the performance gain that would come through by not having the extra ports being carried forward, would come only when you remove the un-wanted ports from the source itself.

Moving an ecommerce site to Amazon Web Services

2014-01-23T21:00:00.000-05:00

This is an interesting one.. There is an e-commerce startup, run by a dear friend. One fine day we were chatting around and he mentioned challenges with his business. Challenges included performance, scaling and cost issues about his hosting services provider.

My instant reaction was, why dont you move to aws.. and it clicked just like that.. We decided that we'd start with the free tier, with his mysql server on amazon RDS and rest of app server functionality on the micro instance. All of which is within purview of the free tier. Using Cloudfront we'd localize the static content, to add to the performance of content delivery. Not now, but eventually, we'd start using ELB - the load balancer for distributing load across the instances we would have.

With something like aws, planning all that is really really that straightforward. You just have to be a bit more technically oriented to think in terms of these things and you are set for good.

We are already thinking about configuring autoscaling for the services, so that running on the micro instance doesnt become a bottleneck for the customers. With that we'd hope to achieve dynamic automatic scaling up and down the infrastructure.

We have already taken the first steps, and I am helping his venture migrate to AWS, one of my dream jobs, to be able to consult around migration / integration of Traditional systems to cloud based systems.

Teradata TPump vs MultiLoad

2013-01-12T00:01:00.000-05:00

Was doing some research around Teradata's load utilities and found some useful info on Teradata forums

MultiLoad
* Loads data to TeraData from a Mainframe or flat file
* Multiple tables can be loaded in the same MultiLoad.
* Up to 20 INSERTS, UPDATES, or DELETES on upt o 5 tables.
* UPSERT is supported
* There can be NO - Unique Secondary Indexes (USI), Triggers, Referential Integrity or Join Indexes.
* Duplicate rows are allowed
* Each Import task can do multiple INSERT, UPDATE and DELETE functions.
* Some Secondary Indexes (NUSI) and RI are allowed
* Locks at the table level
* Block Level transferring of Data.

TPump

* Loads data to TeraData from Mainframe or flat file
* Processes INSERTS, UPDATES or DELETES
* Tables are usually populated.
* Can have Secondary Indexes and RI on tables.
* Does not support MULTI-SET tables.
* Locks at the row hash level
* It uses time based checkpoints not count based.If you are using an OLTP the TPUMPs trickle or continuous loads to populated tables. It acts like a water faucet (tap), that is it can be turned up and load millions of rows or at peak periods tuned down to trickle ffed into the tables.Generally MultiLoad performs better for large bulk loads because of the 64k block loading of data and TPump works better on Low volume changes.

General Writeup
Multiload performs better in almost all cases.

The only time TPump's performance approaches Multiload is when you are updating a very small percentage of the rows. Tpump could probably beat Multiload if you had a very small number of rows in that Multiload has to log on to one session per AMP, whereas Tpump sessions can be controlled. So, with a very few number of rows, the overhead of Multiload may make it slower than TPump.

Multiload performs better because it sends the data from the host to the DBMS more efficiently (in block mode; with no embedded commands). Tpump sends the data as part of a statement (exec macro statement). Tpump allows you to pack statements together to gain more efficiency, but it's still doesn't send the data as efficiently as Multiload.

The second reason that Multiload is faster is that it then applies the updates in block mode. So, if you have multiple updates destined for the same data block, they will get applied with one physical write of the data block. Tpump will need to write the data block once for each update.

The advantage that Tpump has over Multiload is that it locks only the rows (actually row hashes) that it's updating whereas Multiload locks the entire table for write while it's updating the data. Because of this, you can run multiple Tpumps against the same table at the same time, whereas you can only run one Multiload against a table at a time.

Since Multiload takes a write lock on the table it's updating during it's APPLY phase (the phase where it actually updates the table), you can only access the table with an access lock (i.e. dirty read). With Tpump, you could access individual rows with a regular read lock. If you tried to do a query that required a read lock on the table during a Tpump, the read lock would end up blocking the Tpump updates until the query finished, so it's still not a great idea to try to run queries requiring a read lock on the table during a Tpump.

Another advantage that Tpump has over Multiload is that there is no Tpump code within the DBMS, so new features are automatically enabled with Tpump, whereas there are a number of features that you can't use with Multiload (USI's, referential integrity, join indices, etc.).

In summary,

Multiload is better for bulk updating especially if done in traditional batch mode.
Tpump is usually better for continuous updating of a table.

Finding Informatica domain name

2013-01-03T21:30:00.004-05:00

Recently I came across a situation where the customer people had provided us with informatica server hostname, but not the domain name, nor the port for domain.

In such a case, we lost quite some time figuring out how to go through the domain configuration. That's when I started thinking about the alternates for finding the domain name information from the system (assuming different access levels)

If the repository database access is available, i.e. you can access the informatica repository database, you can use the following query to get the domain name out.

select pos_name
from PO_DOMAINSERVICECONFIG

This sql will need to be run in the schema where the domain repository has been created.

On another approach, if the database access is not there, and the informatica server access is available, another file, domains.infa in the $PM_HOME equivalent directory will be able to provide information on domain name/port etc..

About the port

Though the installations process allows customization of the ports for domain, many installations keep the default as is. In any case, a simple telnet to the host on the suspicious port will confirm whether the port is open or not.

In my example situation, it turned out to be the default 6005.

hadoop/hive with tableu

2012-06-13T17:47:00.001-04:00

It was in 2010 that I had the first taste of hadoop/hive. Back then I was still using hadoop 0.20 and was doing a proof of concept for a customer, who wanted to see if hadoop can be a solution for their problems.

Since then, I have been reading up and following the changes in the hadoop world, and tweaking things here and there with the home installation. Today, I tried to mount hive on hadoop (without hbase, with hbase will be the next experiment) and see how can I get it playing nicely.

The experience is awesome as usual, and it reinforces my belief in the fact that hadoop ecosystem has a huge role to play in the computing industry of tomorrow.

The analytical capabilities of the volumes of data managed by the hadoop kind of system are ever increasing, therefore the interest from many instant BI players to provide access to the data behind hadoop.

One such player is the instant dashboard tool - Tableu. They have announced that Tableu 7 will be able to read data directly from hive environments.

In real life it was a bit of a challenge, but whats the fun if there is no challenge. In a nutshell, it does work. No doubt. However, the kind of configuration that is required and administration can be tricky.

1. You have to install the hive driver (available from their website - http://www.tableausoftware.com/support/drivers)

2. you have to launch hive in a particular way, as a service. (hive --service hiverserver). Also, hive on a pseudo cluster only allows one user connected (since the metadata store is single user on Derby). as a result, if you are using Tableu connectivity, nothing else can access hive, not even a command line inerface.

3. Remember that each addition/change to the data set on tableu interface triggers a map-reduce job on the hive cluster/server. And that, hadoop/hive are not really meant to be fast responsive systems. Therefore, expect high delays in fulfilling your drag and drop requests.

4. There might/will be additional troubles in aggregating certain types of data, since the data types on hive might not be additive in the same way as front end expects them to be.

All in all, it wins me in the ease of use provided for accessing the data behind the hadoop environment, however, there are faster ways that already exist to achieve the same result.

my experiments with technology

Error installing pytorch - OSError: [Errno 2] No such file or directory

install google cloud api tools on ubuntu / debian

What you need before Starting

Installation steps

Step 1: Add `gcloud` CLI distribution

Step 2: Acquire the public key

Step 3: Install `gcloud`

Step 4: Initialize the `gcloud` CLI

Mounting Blob containers from Azure using blobfuse2

Apache Spark Interview Questions

Handling data with changing schema on Hadoop

Hadoop - Small Files vs Big Files

Eclipse - installing Scala plugin manually?

Links to free big-data-sets

Teradata Data type abbreviation - described

Hadoop Meetup on the sidelines of Strata Hadoop Conference - Part 2

Hadoop Meetup on the sidelines of Strata Hadoop Conference - Part 1

Data Sets

Weighted trust graph for authentication

Hackathon - Fintechathon

Playing with R

Can we rollback the Truncate, Drop, and Delete command deleted data?

Difference between Char, Varchar, and nVarchar data type in SQL Server

Welcoming Microsoft world onto my blog...

Programmatic interaction with Facebook data

Proud - Cloudera Certified Developer for Apache Hadoop

Removing ports from an existing mapping - Do's and Dont's

Moving an ecommerce site to Amazon Web Services

Teradata TPump vs MultiLoad

Finding Informatica domain name

hadoop/hive with tableu

my experiments with technology

Error installing pytorch - OSError: [Errno 2] No such file or directory

install google cloud api tools on ubuntu / debian

What you need before Starting

Installation steps

Step 1: Add gcloud CLI distribution

Step 2: Acquire the public key

Step 3: Install gcloud

Step 4: Initialize the gcloud CLI

Mounting Blob containers from Azure using blobfuse2

Apache Spark Interview Questions

Handling data with changing schema on Hadoop

Hadoop - Small Files vs Big Files

Eclipse - installing Scala plugin manually?

Links to free big-data-sets

Teradata Data type abbreviation - described

Hadoop Meetup on the sidelines of Strata Hadoop Conference - Part 2

Hadoop Meetup on the sidelines of Strata Hadoop Conference - Part 1

Data Sets

Weighted trust graph for authentication

Hackathon - Fintechathon

Playing with R

Can we rollback the Truncate, Drop, and Delete command deleted data?

Difference between Char, Varchar, and nVarchar data type in SQL Server

Welcoming Microsoft world onto my blog...

Programmatic interaction with Facebook data

Proud - Cloudera Certified Developer for Apache Hadoop

Removing ports from an existing mapping - Do's and Dont's

Moving an ecommerce site to Amazon Web Services

Teradata TPump vs MultiLoad

Finding Informatica domain name

hadoop/hive with tableu

Step 1: Add `gcloud` CLI distribution

Step 3: Install `gcloud`

Step 4: Initialize the `gcloud` CLI