Infor ETL

Hello,

Could i have a couple questions confirmed?

1) The license tool selects ALL variations since the last time it ran? in the case of MITTRA, there is around 10 updates for the same key within minutes. Assuming these are all read.

2) For each of those read, it actually goes to the upsert and does the update for every one of these. The final one is the last update?

3) Are we able to add logic that would a) compare and only update the last variation for a key, or b) do the update and if failed with not found, execute the insert? I am not sure that would be within the license for our use.

Thank you in advance

Find more posts tagged with

Accepted answers

tdouglass

Thank you. I think you explained it. I know the Infor ETL was to be used for getting data out of M3 and staging it to non M3 landing. I think it is just if used for other transformation type logic, Infor wouldn't support that.

All comments

Bojan Rafajloski

Hi TD, the ETL Client has documentation how to create a transformation using the 'QueryAll' function to perform the UPSERT logic with the StagingUpsert step: Creating a new Transformation

To answer your questions:

1) Yes, it will read from Data Lake every variation.

2) Yes, it will perform the upsert logic based on the variation number. Update existing records only if the variation is higher. Inserts records if they do not exist in the table.

3) That is what the StagingUpsert step already does. Please let me know if I misunderstood your question.

tdouglass

Bojan, researching Pentaho (and also my past experience), sometimes there can be better performance with only doing the update, then filtering for errors if record doesn't exist and then doing the insert in the failure path. We are just trying to do some performance issue resolution.
Also, our mittra has sometimes 100+ for the same key and wasn't sure if better performance by reading each record and comparing the variatoin number and only doing the insert/update for the greatest variation in that group of data.

tdouglass

This is the current pentaho feedback i have found. What is the license restrictions of the tool?

The Insert/Update operator is very slow, usually is more efficient to separate the rows to insert from the rows to update, and perform the two operations separately, in two separate transformations to avoid locking the table by inserting and updating rows in two separate operations at the same time.

Bojan Rafajloski

Currently there are no plans to change the logic of the Infor provided StagingUpsert step in the ETL Client. You could try customizing your transformations to use what the Pentaho tool provides and build your own logic, but Infor would be limited in supporting issues with the customization.

Could you explain which license restrictions are you referring too? Infor ETL Client for Data Lake is licensed as Enterprise License. Please contact your Infor account manager to discuss the details of the license.

Copyright © 2025 Infor. All rights reserved.