Max Query Count setting

[former_member182065]

Hello colleagues,

Probably you many know there is Max Query Count setting in MII.

Does any of you experience performance concerns by dealing with large number of query count? or does any of you have a recommendation for how to set max query count, e.g. not exceeding memory size to avoid memory swapping?

my customer want to increase this max count due to fluctuating record count. but they have concerns on such disadvantage in performance.

any of your experiences are welcome. Thank you,

Shiroh Kinoshita, SAP Japan

[jcgood25]

Shiroh,

Good to hear from you.

All queries have the RowCount setting, but we typically discourage people from just setting it to a high number. In some cases the data servers will limit this number to a max (something like 250000) but that doesn't mean that from a memory standpoint, or a customer patience level (especially in the browser) would ever get to that number of records.

Where is it that you are seeing the record count fluctuate, applets in the browser, query actions in a transaction?

What is it that makes the customer want to increase this to a high number?

Regards,

Jeremy

[former_member182065]

Jeremy,

Good to hear from you too. I will welcome Kevin next week.

In the meantime, the customer wants to increase number just because it exceeded expected level, which is around 150, that is ERP's shipment documents. and would like to increase something like 200 or 10000, etc. They deal with the query in transaction. But, when I see the previous message about material record testing from ECC IDES, that looks not so much big issue.

But I will recommend the customer set the max count lower level and proceed with the test.

Thanks for your advice.

Sincerely, Shiroh Kinoshita - SAPJ

[Former Member]

Actually this brings up a question I've always had.

How does MII actually handle max row count? Does it send the max rowcount as part of the query? or does it limit it as the records are returned. Or does it depend on the data source? I would think it would make sense to limit the query in the first place but depends on the type of query. For example if your are looking at historian data and doing a time based query of raw values.

[jcgood25]

Doug - depends upon the data server type.

For JDBC, yes it does set the row count through the connector, but if more records come back it won't cut them off (that's why SET ROWCOUNT 0 in a fixed query can be used, even if the RowCount for the template is 100)

The XacuteConnector when calling a TRX will chop the records off at the RowCount number in the Xacute Query, but of course can't prevent the work done in the trx from trying to send back more.

UDC type - probably in a similar capacity as JDBC, but not sure because of the varying ways API and underlying standards like OPC behave.

[Former Member]

i just did a test, i got all materials from ecc6_ides to mii and there are about 2800 rows records.seems it's ok, 5 seconds

[former_member182065]

Thank you very much. I have not noticed using ECC IDES system. that sounds very good!

Shiroh Kinoshita