.ADI index gets corrupted when deleting large amounts of records

[Former Member]

We have a rather large table (93 fields, 25 indexes), which I can consistently corrupt an index on by deleting a large amount of records.

I currently have a table of 459938 records.

If I attempt to delete 426744 of these using SQL (delete from TABLE1 where FIELD1>='8'), I get the message "poQuery: Error 7200: AQE Error: State = HY000; NativeError = 7017; [iAnywhere Solutions][Advantage SQL][ASA] Error 7017: Corrupt .ADI, .CDX or .IDX index."

Running a further select statement I can see only 2825 records were removed before the error occurred. Retrying the delete query results in one further record being deleted every time the query is run before getting the error again.

Deleting and recreating the indexes will fix the problem until the delete query is run again.

The problem can be reproduced on various servers using ADS 11.10.0.10 but I am currently using WindowsServer2008R2.

I can also reproduce a very similar problem by simply highlighting all the records in Arc32 and pressing the "Delete current or selected records" button.

[HakanHaslaman]

Thanks for confirming the issue. The engineering are working on it, but currently we cannot commit a time frame for the fix.

[HakanHaslaman]

Kev, please try it with the version 11.10.0.24

[Former Member]

Hi Hakan. Sorry for the delay, we have now updated to ADS 11.10.0.24 but we still get the same problem.

[HakanHaslaman]

Please send me a re-creation with a test database and test records (with a step by steps repro description) to analyze this issue.

[Former Member]

Thanks. Shall I send it to the "sapnetwork@sap.com" address?

[HakanHaslaman]

Please give us a download link.

[Former Member]

That's fine Hakan. Where would you like me to send the download link? I'd rather not put it up publicly.

[HakanHaslaman]

send me over the direct message

[HakanHaslaman]

Thanks for the link, I will forward this behavior to the engineering and if I get any news, I will let you asap know about it.

[HakanHaslaman]

Short update: we are able to reproduce this issue, so the engineering started to investigate on it. Will get back to you asap.

[HakanHaslaman]

The engineering narrowed down the issue and found that index WHORDDTD is causing the problem. Please try in index WHORDDTD to currently remove DESCEND () from DETDREQ;DESCEND(DETMARGIN) and modify the index to DETDREQ;DETMARGIN.

[Former Member]

That does seem to be the cause of the problem!

Is this something that will be fixed? We have software that relies on the index using DESCEND().

Thanks for your efforts so far.