dbanalys -> no way to do this in 4GL?

[palthe]

Hi,

I was wondering; in which table (system) is the dbanalys information stored? I can't seem to find any metadata table and this is by far the fastest way to get tablescan info.

Another question regarding the dbanalys command; is the number of records accurate information or index information?

 

[rstanciu]

you can do a proutil db-name -C updatevst
and check these informations. There is a entry in the OpenEdge standard documentation
about "Virtual System Tables" to got tables descriptions.

 

[TomBascom]

Hi,

I was wondering; in which table (system) is the dbanalys information stored? I can't seem to find any metadata table and this is by far the fastest way to get tablescan info.

Another question regarding the dbanalys command; is the number of records accurate information or index information?

No, this information is not stored. (IMHO it should be though.) VSTs do not contain record count, min/max/mean, scatter or fragmentation data nor is index utilization info persisted. You must run dbanalys and parse the output file.

The data is accurate if it is run offline. If it is run with the db up then the numbers are obviously subject to the whims of what may have happened internally while the report was runnng -- but for most purposes they are more than good enough.

 

[palthe]

No, this information is not stored. (IMHO it should be though.) VSTs do not contain record count, min/max/mean, scatter or fragmentation data nor is index utilization info persisted. You must run dbanalys and parse the output file.

The data is accurate if it is run offline. If it is run with the db up then the numbers are obviously subject to the whims of what may have happened internally while the report was runnng -- but for most purposes they are more than good enough.

Yes that was what I was afraid of... (no storage of metadata). I already made a script that can parse the output but it feels like a workaround (which it really is). I know have to choose between somewhat inaccurate information which is gathered with a workaround, or a really slow tablescan.

Thanks for the answer, I now know I'm not missing something here.