Replication Set date stamps

We are setting up some replications to load a data warehouse, and a few concerns with the process were raised. Hoping to get some advice on a few questions:

1. Is there a way to add a timestamp to the export filename? Some concern was raised about the potential of overwriting a previous file if the files aren't pulled from the SFTP site for some reason.

2. Has anyone pushed replication set files to an Azure blob site directly?

Find more posts tagged with

Accepted answers

All comments

robert-nevinger

Robert:

I don't believe you can add a timestamp to the ReplicaitonSet name, but you have some options. First, you can use the Replication component parameters that allow you to push ReplicationSets to folders outside Administration Console, or your own dedicated SFTP server.

If you are not familiar with this, navigate to Admin Console > Other > System Configuration Parameters > By Preferred Category
Enter Replication into search. With the results you will see 2 tabs: Defined & Available.
Defined means they were created and associated with the Replication component.
Available means they are not created, and you can create them with a right click, and set a value.
Please note that you can define new values for defined parameters.

If you opt for a dedicated FTP server, say one you create in Azure server or AWS EC2, you might have a server configuration option that always adds a data stamp. If you opt to keep the FTP server as is, you could create folders for the Replication Sets e.g. /ReplicationSets/new, then have a FileChannel monitor that folder and triggers an IPA that picks up the file as _inputData and renames the file on the same sftp server, but publishes to say /ReplicationSets/datestamped.

As far as Azure blob, I'm an AWS guy so I publish directly to AWS S3 because Infor supports AWS keys. From there I have a Lambda function push to MongDB, because I opted to have the ReplicationSet publish as JSON.

If you are using Birst Analytics, I don't think you can repurpose the Replication Sets bound for DataLake, you will need to create new RepSets and change the destination to FTP.

You also have to consider full or incremental replication. If you are dong full replication and have large data sets, do your best to keep your IPA streamlined or will wind up with huge workunits. If they get too big you get memory alerts. it looks bad but the flow still completes.

Dr. Awesome

The Replication configuration parameter export.file.name.versioning can be used to create unique file names. It isn't a date that is appended, but a unique number:

E.g.

BackOrderCode_CurrentData520.json

The unique number here is 520. Starting with 0, each subsequent file will increment and append a number to that unique number.

BackOrderCode_CurrentData521.json

BackOrderCode_CurrentData522.json

BackOrderCode_CurrentData523.json

etc.

robert-canham

That looks like an easy way to do it. So this would extend the numbering it does when breaking the output to chunks and just doesn't reset to 0 after the run is done?

Dr. Awesome

It uses the replication set variation id for the unique number -- this will change with each run of a Replication. So unless you delete a Replication and recreate it completely the file name should stay unique and incrementing.

robert-nevinger

Dr. Awesome:

Of course I would be wordsmithing with the master. Does this parameter add a number if chunk size is exceeded, and a replication results in multiple files ?

Copyright © 2025 Infor. All rights reserved.