• 1 reply
  • Latest Post - ‏2013-02-20T13:36:55Z by smithha
12 Posts

Pinned topic Ideas needed on how to execute IA HTTP API commands from DataStage.

‏2013-02-20T00:02:38Z |
I am working on automating my Information Analyzer process. I got everything working using the curl command in Linux. My infrastructure team would like me to use DataStage jobs instead of running scripts on the server. Does anyone know if it is possible and how to trigger IA processes through HTTP API from DataStage. Will any of the stages in DataStage support running the IA urls that work with HTTP API? Maybe there is another way without triggering Linux commands to get the data through HTTP API. I was trying to find some way without calling an external command from the operating system.

I have the process working with the curl command to connect using HTTP API. Basically I am doing the following steps:
1) Creating a backup of the current project, prior to making modifications.
2) Creating a new virtual table to limit the records I am checking.
3) Modifying the existing rule to bind to the new virtual table.
4) Executing the rule.
5) Extracting the invalid records.
6) Converting the extracted XML to a spreadsheet XML format. This allows the users to directly open the file in Excel.
7) Emailing the output.
Updated on 2013-02-20T13:36:55Z at 2013-02-20T13:36:55Z by smithha
  • smithha
    161 Posts

    Re: Ideas needed on how to execute IA HTTP API commands from DataStage.

    The easiest approach is probably to build a DataStage job sequence (which hopefully would satisfy your infrastructure team since they can schedule it).
    You can use sequence stages (ExecCommand) to invoke the IA CLI for each of the steps you need. There are also sequence stages (Notification) to handle email functions. And you can also include sequence stages to check whether the rule has finished execution before performing extract etc.

    Another approach (if you are on version 8.7 or higher) would be to use the added RuleStage to run the IA rules directly instream in DataStage. You'd need to change your approach to use a standard DS connector stage to perform the data extract (virtual table selection), then run the rules and output to whatever format you want. You'd still need to put into a job sequence to handle email notification, though.