An Overview of DB2 Federation

DB2 Federation
DB2 Federation

Data analytics has changed where data is no longer manageable in relational databases only. Data is flowing from various sources which are not of the same format. This means it is not possible to store all data in the same repository. Some are best suited for storing in relational databases, others for Apache Hadoop while others are best suited for NoSQL databases.

During data analyzing, so much time is taken in trying to bring the distributed data together instead of obtaining insights. Db2 Federation has come to the rescue of data analysts. Federation concept in db2 eliminates the need for storing data in different repositories and reduces the hassle of getting insights.

What is DB2 Federation?

DB2 federation is a data integration technology that permits remote database objects to be accessed as local DB2 database objects. This technology connects multiple databases and makes them appear like one database.

How does DB2 federation work?

Federation allows you to access all of your data that is on multiple distributed databases using a single query. When implemented in an organization, this technology can be used to access data that is on any of the organization’s Db2, whether local or in the cloud.

Why use DB2 federation?

So, why should you use the federation? This concept brings data of all formats into one virtual source. With data being retrieved from one virtual source, analyzing it becomes cost-effective and efficient.

What are its primary use cases for DB2 federation?

Merging of various sources of data

DB2 federation facilitates consolidating of data from sources data local and cloud to form one virtual data source. This eliminates the process of migrating data which can be expensive and troublesome.

Increase the capacity of a repository beyond the fixed limits

Physical storage capacity is bound to have a limit which is one reason you may find an organization has distributed its data in various repositories. With federation, the storage is virtual and therefore doesn’t have any limit. This technology can greatly help you if your physical dataset is running low on space.

Linking up to Db2 Warehouse on Cloud

People who use Db2 products can federate data from Db2 on Cloud and Db2 Warehouse on the Cloud. This will give them a joint interface where they can access, add, query, and analyze data without encountering the complex ETL processes. Better still, no additional code will be required to execute all these processes. This makes it easy for people with the low technical know-how to use these products smoothly.

Split data across different servers

At times, you might choose to partition your data. With federation integration technology, partitioned data can be queried with a unified interface. Federation allows you to better balance your workloads, scale precise parts of an app, and create micro-services that work harmoniously.

Generally, db2 federation makes it access data by bringing it together into a single virtual source. This brings about cost and time-saving benefits. When you want to analyze data, you can get insights immediately instead of spending a lot of time querying through repositories.

Related References

DataStage – Netezza Connector Action Column

Over the years have occasionally use the action column feature, however, the last month or so I have found myself using it quite a lot. This is especially true in relation to the tea set and not just in relation to the change capture stage.

The first thing you need to know is, if you want to prevent getting the ‘no action column found’ notice on the target stage, need to ensure that the action column has been coded to be a single character field char (1). Otherwise, the Netezza connector stage will not recognize your field as an action column.

While most developers will commonly work with the action column feature in relation to the change capture stage, it can also be very useful if you have created a field from one or more inputs to tell you what behavior the row requires. I have found that this approach can be very useful and efficient under the right circumstances.

Example Pattern for Action Column Using Multiple Source Selects
Example Pattern for Action Column Using Multiple Source Selects

Action column configuration example

Action Column Field Type
Action Column Field Type

 Change Code Values Mapping To Action Column

  • Here’s a quick reference table to provide the interpretation of the change type code to the actual one character action column value to which it will need to be interpreted.

Change Code Type

Change Type Code

Action Column Value

Copy (Data Without Changes)

0

No
value for this Change Type

Insert

1

I

Delete

2

D

Update

3

U

Example Transformer Stage, Derivation

  •  Here is a quick transformer stage derivation coding example to take advantage of the action call capabilities. If you haven’t already handled the removal of the copy rows, you may also want to add a constraint.
  • The combination I most frequently find myself using is the insert and update combination.
if Lnk_Out_To_Tfm.change_code=1 then ‘I’

Else if Lnk_Out_To_Tfm.change_code=2 then ‘D’

Else if Lnk_Out_To_Tfm.change_code=3 then ‘U’

Related References

Home > InfoSphere Information Server 11.7.0 > InfoSphere DataStage and QualityStage > Developing parallel jobs > Introduction to InfoSphere DataStage Balanced Optimization > Job design considerations  > Specific considerations for the Netezza connector

Netezza / PureData – List of Views against a table

PureData Powered by Netezza
PureData Powered by Netezza

I have found myself using this simple, but useful SQL time in recent weeks to research different issues and to help with impact analysis.  So, I thought I would post it while I’m thinking about it.  It just gives a list of views using a table, which can be handy to know.  This SQL is simple and could be converted to an equi-join.  I used the like statement mostly because I sometimes want to know if there are other views a similar nature in the same family (by naming convention) of tables.

Select All Fields From The _V_View

This is the simplest form of this SQL to views, which a table.

Select * from _v_view

 

where DEFINITION like ‘%<<TABLE_NAME>>%’ ;

Select Minimal Fields From The _V_View

This is the version of the SQL, which I normally use, to list the views, which use a table.

Select VIEWNAME, OWNER from _v_view

 

where DEFINITION like ‘%<<TABLE_NAME>>%’;

Related References

Netezza / Puredata – How to replace or trim CHAR(0) is NULL characters in a field

PureData Powered by Netezza
PureData Powered by Netezza

Occasionally, one runs into the problem of hidden field values breaking join criteria.  I have had to clean up bad archive and conversion data with hidden characters serval times over the last couple of weeks, so, I thought I might as well capture this note for future use.

I tried the Replace command which is prevalent for Netezza answers to this issue on the web, but my client’s version does not support that command.  So, I needed to use the Translate command instead to accomplish it.  It took a couple of searches of the usual bad actors to find the character causing the issue, which on this day was chr(0).  Here is a quick mockup of the command I used to solve this issue.

Example Select Statement

Here is a quick example select SQL to identify problem rows.

SELECT TRANSLATE(F.BLOGTYPE_CODE, CHR(0), ”) AS BLOGTYPE_CODE, BT.BLOG_TYP_ID, LENGTH(BT.BLOG_TYP_ID) AS LNGTH_BT, LENGTH(F.BLOGTYPE_CODE) AS LNGTH_ BLOGTYPE

FROM  BLOGS_TBL F,  BLOG_TYPES BT WHERE TRANSLATE(F.BLOGTYPE_CODE, CHR(0), ”) =  BT.BLOG_TYP_ID AND LENGTH(BT.BLOG_TYP_ID) <>Length(LENGTH(F.BLOGTYPE_CODE) ;

 

Example Update Statement

Here is a quick shell update statement to remove the Char(0) characters from the problem field.

Update <<Your Table Name>> A

Set A.<<Your Field Name>> = TRANSLATE(A.<<Your FieldName>>, CHR(0), ”)

where length(A.<<Your Field Name>>) <> Length(A.<<Your FieldName>>) And << Additional criteria>>;

 

 

 

SQL Server – how to know when a stored procedure ran last

Microsoft SQL Server 2017
Microsoft SQL Server 2017

This week I needed to know if a stored procedure was running when expected during our batch.  So, here is a quick couple of SQL to answer the question:

When a Stored Procedure was run last

This version of the SQL gives the date for the last time the Stored Procure was run:

select distinct   top 1     s.last_execution_time

from  sys.dm_exec_query_stats s

cross apply sys.dm_exec_query_plan (s.plan_handle) p

where  object_name(p.objectid, db_id(‘<<DATABASE_NAME>>’)) = ‘<<STORED_PROCEDURE_NAME>>’

Order by s.last_execution_time desc

Get a list of when Stored Procedure has been run

This version of the SQL provides a list of dates of when Stored Procure has been run:

select distinct   s.last_execution_time

from  sys.dm_exec_query_stats s

cross apply sys.dm_exec_query_plan (s.plan_handle) p

where object_name(p.objectid, db_id(‘<<DATABASE_NAME>>’)) = ‘<<STORED_PROCEDURE_NAME>>’

Order by s.last_execution_time desc

 

 

Netezza / PureData – How to Substring on a Character

PureData Powered by Netezza
PureData Powered by Netezza

 

I had a reason this week to perform a substring on a character in Netezza this week, something I have not had a need to do before.  The process was not as straightforward as I would have thought, since the command is explained as a static position command, and the IBM documentation, honestly, wasn’t much help.  Knowing full well, that text strings are variable having to provide a static position is not terribly useful in and of itself.  So, we need to use an expression to make the substring command flexible and dynamic.

I did get it work the way I needed, but it took two commands to make it happen:

  • The First was the ’instr’ command to identify the field and character I wanted to substring on: instr(<<FIELD_NAME>>,’~’) as This provides the position number of the tilde (~).
  • The second was the ‘substr’ command in which I embedded the ‘instr’ command: substr(<<FIELD_NAME>>,0,instr(<<FIELD_NAME>>,’~’) )

This worked nicely for what I needed, which was to pick out a file name from the beginning of a string, which was delimited with a tilde (~)

Substring on a Character Command Format

  • This format example starts with position zero (0) as position 1 of substring command and goes to the first tilde (~) as position 2 of the substring command.
Select  <<FIELD_NAME>>

, instr(<>,’~’) as pos2

, substr(<<FIELD_NAME>>,0,instr(<<FIELD_NAME>>,’~’) ) as Results

From <<Table_Name>>

where  <<Where_Clause>>;

 

 

Related references

IBM Knowledge Center, Home, PureData System for Analytics 7.2.1, IBM Netezza database user documentation, Netezza SQL basics, Netezza SQL extensions, Character string functions

IBM Knowledge Center, Home PureData System for Analytics 7.0.3, IBM Netezza Database User’s Guide, Netezza SQL basics, Netezza SQL extensions, Character functions

Oracle TO_CHAR to SQL Server CONVERT Equivalents to change Date to String

Transact SQL (T-SQL)
Transact SQL (T-SQL)

When it comes to SQL I tend to lean on the SQL I have used the most over the years, which is Oracle.  Today was no exception, I found myself trying to use the TO_CHAR command in SQL Server to format a date, which of course does not work. So, after a little thought, here are some examples of how you can the SQL Server Convert Command the achieve the equivalent result.

Example SQL Server Date Conversion SQL
Example SQL Server Date Conversion SQL

Example SQL Server Date Conversion SQL Code

This SQL of examples runs, as is, no from table required.

 

Select

CONVERT(VARCHAR(10), GETDATE(), 20) as
‘YYYY-MM-DD’

,CONVERT(VARCHAR(19), GETDATE(), 20) as ‘YYYY-MM-DD HH24:MI:SS’

,CONVERT(VARCHAR(8), GETDATE(), 112) as YYYYMMDD

,CONVERT(VARCHAR(6), GETDATE(), 112) as YYYYMM

,CONVERT(VARCHAR(12), DATEPART(YEAR, GETDATE()))+ RIGHT(‘0’+CAST(MONTH(GETDATE()) AS VARCHAR(2)),2)
as
YYYYMM_Method_2

,CONVERT(VARCHAR(4), GETDATE(), 12) as YYMM

,CONVERT(VARCHAR(4), GETDATE(), 112) as YYYY

,CONVERT(VARCHAR(4), DATEPART(YEAR, GETDATE())) as YYYY_Method_2

,CONVERT(VARCHAR(4), YEAR(GETDATE())) as YYYY_Method_3

,RIGHT(‘0’+CAST(MONTH(GETDATE()) AS VARCHAR(2)),2) as Two_Digit_Month

,SUBSTRING(ltrim(CONVERT(VARCHAR(4), GETDATE(), 12)),3,2) as Two_Digit_Month_2

,CONVERT(VARCHAR(10), GETDATE(), 111) as ‘YYYY/MM/DD’

,CONVERT(VARCHAR(5), GETDATE(), 8) as ‘HH24:MI’

,CONVERT(VARCHAR(8), GETDATE(), 8) ‘HH24:MI:SS’

 

Map TO_CHAR formats to SQL Server

You can map an Oracle TO_CHAR formats to SQL Server alternative commands as follows:

TO_CHAR
String

VARCHAR
Length

SQL
Server Convert Style

YYYY-MM-DD

VARCHAR(10)

20,
21, 120, 121, 126 and 127

YYYY-MM-DD
HH24:MI:SS

VARCHAR(19)

20,
21, 120 and 121

YYYYMMDD

VARCHAR(8)

112

YYYYMM

VARCHAR(6)

112

YYMM

VARCHAR(4)

12

YYYY

VARCHAR(4)

112

MM

VARCHAR(2)

12

YYYY/MM/DD

VARCHAR(10)

111

HH24:MI

VARCHAR(5)

8,
108, 14 and 114

HH24:MI:SS

VARCHAR(8)

8,
108, 14 and 114

Translating the formats commands

Here are some example of translating the formats commands.

Format

SQL
Server

YYYY-MM-DD

CONVERT(VARCHAR(10),
GETDATE(), 20)

YYYY-MM-DD
HH24:MI:SS

CONVERT(VARCHAR(19),
GETDATE(), 20)

YYYYMMDD

CONVERT(VARCHAR(8),
GETDATE(), 112)

YYYYMM

CONVERT(VARCHAR(6),
GETDATE(), 112)

YYMM

CONVERT(VARCHAR(4),
GETDATE(), 12)

YYYY

CONVERT(VARCHAR(4),
GETDATE(), 112)

YYYY

CONVERT(VARCHAR(4),
DATEPART(YEAR, GETDATE()))

YYYY

CONVERT(VARCHAR(4),
YEAR(GETDATE()))

MM

RIGHT(‘0’+CAST(MONTH(GETDATE())
AS VARCHAR(2)),2)

MM

SUBSTRING(ltrim(CONVERT(VARCHAR(4),
GETDATE(), 12)),3,2)

YYYY/MM/DD

CONVERT(VARCHAR(10),
GETDATE(), 111)

HH24:MI

CONVERT(VARCHAR(5),
GETDATE(), 8)

HH24:MI:SS

CONVERT(VARCHAR(8),
GETDATE(), 8)

Related Reference

Microsoft Docs, SQL, T-SQL Functions, GETDATE (Transact-SQL)

Microsoft Docs, SQL, T-SQL Functions, Date and Time Data Types and Functions (Transact-SQL)

Microsoft Docs, SQL, T-SQL Functions, DATEPART (Transact-SQL)