Professional Documents
Culture Documents
7) Create new informatica folder and unix folders with standard naming convention
15) Co-ordinate with development team to setup the automated ftp for source files
Design documents
The design documents such as the SRS and the SRDS should be updated and all the
details should be present, all hyperlinks should work and the parts not applicable
should be clearly highlighted. This will help to save time and effort to review the
entire document since the SRS and the SRDS are detailed documents, which need a
considerable time and effort to review
• Transformation Specifications
• Capacity Plan
• Test Plan
• Impact Analysis Document
• High Level Logical Data Model
GE Confidential
Transformation Specifications
The transformation specifications should contain details about all the transformations
that have been used in the mappings the transformation logic should be explained on
a high level. This should also have information about the source and the target type
viz the source or the target is a flat file, or RDBM system.
Capacity Plan
Test Plan
The Unit Test Plan and the System Test Plan should be prepared highlighting all the
test cases and also it should have information that whether the entire module passed
all the tests in the development environment. Any sort of extraordinary errors
encountered should be mentioned.
GE Confidential
output ports) in lookups, the "O" option should be
9. Hard coding in sql override (in source qualifier or Lookup sql override)
should be avoided or if possible should be made generic. e.g. The fiscal
month,year,date related logic should be implemented dynamically instead of
hard coding if possible.
GE Confidential
should be closer to source, if you are updating only or inserting only, use
option of treat rows as update/insert as required rather than using update
strategy and treating rows as data driven. If you are using variables in
expressions, the sequence of ports should be input, variable, then output
ports. If possible, use sql override in lookups to bring only required data
from database to cache using filter conditions.
19. If the source are files, the details of source files should be provide in MD120
like from which system it is going to come, what is the frequence, who is
contact person etc.
20. As decided, the migration from development to test, test to preprod is done
by changing the existing connections details to required environment
connection details.eg.INV_DEV_INS will be change to INV_TEST_INS and
internally it will be pointed to test environment connections.
21. No presession or post session command are allowed. That should be step in
cronacle job chain. It makes easier to debug the failures.
24. DB link are not allowed to use in source qualifier sql override.
25. Tracing level for all the transformation should be NORMAL.
26. If source data is having Chinese /Japanese characters, the code page for the
oracle database connections should be UTF- 8
27. If the source data have Unicode character, to handle those Unicode
characters, it’s mandatory to mention to use UTF-8 code page for creation of
Teradata Relational Connections. If mload or fastload is being used, the
output file properties to be set to have UTF-8 code page.
28. The step of Generation of Reject records log and sending in email should be
added at the end of Cronacle Job chain.
29. The source and target definition changes should be done by importing the
same from database instead of making changes in definitions MANUALLY in
Informatica designer.
GE Confidential
Repository should be clean i.e. should have only mapping, sessions and workflow
those need to be migrated.
Reference 1. http://gemsbidev.med.ge.com/opsmanual/
2. http://uswaubus02medge.med.ge.com/webhouse/reports/BI_Life_Cycle.htm
Informatica Standards
GE Confidential
Transformation
Database XXX_<Database Name>_<Schema Name>
Connections
ODBC Connection <Schema_Name/ Data_Base_Name>
Name For Oracle use Schema Name e.g. INDLOAD
For Teradata use Data Base Name: e.g. SRC_ETL_TARGET
NOTE* Do not use any suffix or prefix to the ORDBC Connection name.
Eg. SRC_ETL_TARGET_1 or SRC_ETL_TARGET_test etc. There should be
only one ODBC Connection for one Database for same Project
GE Confidential
4.1 Naming convention for Informatica Objects (Refer Naming
Convention for Informatica Objects section in this Document).
1) SQL Override: No SQL Override should be used in the Source Qualifier unless
used to join multiple tables. It should not be overwritten to pass any filter. In
such scenario just mention the filter in the filter section.
2) Hard Coded DB Name: There should not be any “hard coded” database name
in any sort of SQL Override. This is applicable to SQL Override used in the SQ
or Lookup
3) Filter Usage: Use Filter in the SQ if possible or as close as possible to SQ.
4) Update Strategy Transformation: While loading to a Teradata target Use
UPSERT /DELETE/UPDATE Logic in session level instead of UPDATE STRETEGY
Transformation.
5) Unconnected Ports: No Ports should connected forward if not used by any
Transformation.
6) Lookup Vs. Unconnected Lookup: If same technicality can be achieved by
both Connected and un-connected Lookup, connected lookup should be used.
7) Usage of Lookup: If it is avoidable using a lookup by putting it into SQ, it is
suggested to do so.
8) Caching: Caching should be enables in Lookup
9) Joiner: Joiner should not be used to join two table from same Database
10) Target Load Order: Multiple”Target Load Order” should not be used in a Single
Mapping
11) Filter Used for Testing: All the filters or customization done in the mapping for
the testing purpose has to be removed.
12) Sequence Generator: While migrating to QA/PREPROD (M3/M4 Review) Initial
value should be set 1.
GE Confidential
(Note* ODBC connection Name should be mentioned in case of External
Loader Connection used)
6) Rest of the options should be as default unless other wise required.
GE Confidential
4.5 Checklist for MD120:
1) All the Mapping, Session, Workflow name, which need to be migrated, should
be mentioned Correctly
2) All the Source & Target Connection should be mentioned Correctly. In case the
source is flat file, the path and file name should be mentioned. During M3
Migration Mention the DB Connection or Path with respect to TEST/QA Server.
During M4 review it should be done for PREPROD Server.
3) Incase of loading to Teradata, Loader Type (MLOAD/FLOAD) and type of load
(Insert, Update, Delete, Upsert, and Trunc-Insert) should be clearly
mentioned for each session.
4) Provide all the Lookup Database Connection Information for each mapping.
5) If any of the CTL file has been created as READ-ONLY, please provide clear
instruction to migrate them into respective environment and keep them as
READ-ONLY. Also provide the CTL File name, current location and instruction
to change the Log On Information, Database Name & Source File Path for
CTL file. Keep a backup of these READ-ONLY CTL Files.
6) Provide a list of the Source/ Target tables expected to present before start
loading in the new environment
7) Provide a list of the Lookup tables expected to present before start loading in
the new environment
8) Provide the list of UNIX Scripts to be migrated, their location, and changes to
be made in the script during migration.
9) Provide instruction to set the Initial Value and Current Value to 1 for all
sequence generators.
Source Files
All source files need to be defined with their server locations, file names, time of
arrival, estimated size.
Time Windows
All scripts to be defined with their scheduled start and scheduled finish times.
Project
Each script should be part of one and only one project. The project name should be
defined using the following nomenclature. <Value
Chain>_<Module>_<Program>_<Project>. All the sub-parts should be in sync with
the End-to-End Inventory.
Notification
For each script, email ids need to be specified for people who will be notified in case
of failure, success and overdue.Contact Details will be used to contact only in case of
emergencies and for Error Mails. BI Mailing List and Func Mailing List will be used for
the purpose of sending delay mails only.
Dependencies
GE Confidential
In case any job is dependant on flag files or job chains, it needs to specified in the
dependencies and also mention the name of the job that creates the flag file
Checklists
Job chain
Dependencies
1. All Events are mentioned and are part of the job chain script provided in
CVS
2. Are all source tables part of the dependent job chains
3. In case of File Events the location of the source as well as the contact
details are mandatory
4. What will be the corrective action in case the input data file doesn't arrive?
Scripts
1. If project specific scripts are used the description should be explain the
purpose of creating such a script as against the usage of the generic
script.
Time Window
1. Check if any of the source tables are being loaded in the time
frame when the job chain is expected to run.
Are all source tables, lookup tables, database objects and dependant flag files
mentioned.
GE Confidential