Recommended number of documents to update at one time via Keyword Updater?

[Liz_Boultinghou]

Hi all!

I have a situation where I need to add a "closed date"keyword to some of our documents so that we can purge them.

We have discovered that there are approximately 12.5 million documents the the KW updater will need to go through in order to add the keyword to the appropriate ones.

I understand that this can be very system intensive. I'm trying to understand HOW system intensive, so I can set up the correct process(s) and schedule them properly.

Is 12.5 million documents doable in a few hours (assuming I choose "global")?  or would it be better for me to break it up and do batches of document types? I will be running a test, but would like to have some inkling of what to expect.  🙂

Thank you!

Liz

Amplify Credit Union

[Trond_Aas_Ander]

Hi Liz.

My company has been using Onbase for 3-4 months now, and my experience with Keyword Updater is mixed. It seems to work much slower than expected, and we have not managed to update high volumes of documents. Neither when using Keyword Update - Global  nor Keyword Update – Doc Type. The processing time per document is too high. If I remember correctly 20 mill docs would take about 30 days.

When we had the need to update a keyword globally the Hyland database team helped us by executing an sql directly in the database. This worked fine, but is not best practice.

My experience is the KW updater is only functional when the update can be limited to a small number of documents, in example a DT containing relatively few documents. The handling time per document is still much higher than what is expected, but a limited number of documents keeps total processing time down.

I am trying to understand if document handle or other parameters or keyword can be used to limit KW update volumes, and will probably do some testing in the coming weeks.

I would be most interested to know how the high volume keyword update worked for you, and what was your chosen strategy.

 Thanks

Trond Andersen, Sparebank1 Insurance, Norway

[Michael_McElyea]

Liz, the best I have achieved after using Keyword Updater is around 5.1 million in approx. 24 hours.

Based on my observations and different trial runs I found the tool worked best when breaking up records in 50,000 line files. Currently I am using CSVs though TXT did not appear to work better or worse. They are "|" delimited though I have found no change in speed using commas.

I am running our jobs on 17 virtual servers each feed 6 files, again 50,000 lines a piece, using one search keyword and one replace keyword. Seem similar to your scenario.

Ultimately, I think performance is largely dependent on the Database health from what I have found thus far and how the file loads are configured. Like Trond described I found no speed difference when using Doc Type specific vs Global.