Question Re: Request ID in Replication Flows in Technology Q&A

Request ID in Replication Flows

tmeyer — Thu, 26 Sep 2024 07:44:03 GMT

Hi,

we want to store in a local table (which is the target of a replication flow) a kind of request id (like in BW). With a data flow, I could achieve this with a calculated column. I can set a date time field for the current timestamp (which can differ on each record because the data flow takes time).

I thought about a custom annotation or something similar. Has anyone an idea?

Best regards,

Tobias

Re: Request ID in Replication Flows

Martin_Kuma — Fri, 27 Sep 2024 08:06:18 GMT

Hi Tobias,

I would create a custom table, with the 'flow ID' and timestamp as keys. 'request ID' should be system-wide unique so I would concat the flow ID and timestamp.

In a task chain:

1. Update the custom table with a transformation flow -> enter the flow ID and time stamp

2. Load the data into a 'load' table

3. use View to move from 'load' table to 'consumption' table and add the 'request' ID using SQL

I see the overhead in tracking / creating the 'request' ID. You will need to have a custom transformation flow for each load. It would also be possible to omit the step 2 and load directly into a 'consumption' table, but you need to use SQL. Python is too slow.

Cheers

Martin

Re: Request ID in Replication Flows

tmeyer — Fri, 27 Sep 2024 08:56:26 GMT

Hi Martin,

thanks for the idea. I think that would not work because the replication flow is streaming the data from the s/4 system every hour and I don't start it with a task chain.

Br,

Tobias

Re: Request ID in Replication Flows

Martin_Kuma — Wed, 02 Oct 2024 10:45:51 GMT

Hi Tobias, so you have a scheduled replication flow which runs every hour? Then schedule it via Task Chain. If you have the replication flow set for permanent / real-time then you will need some DB trigger based function, else you can use the data flow, but with my described way to fetch the 'request ID' from a custom table. Br, Martin