antoine_mpo
Reputable Poster
Hi List,
We created a custom interactive application with a form, based on a custom view. This view is based on the F41021 (Item Location File) and the F4108 (Lot Master), including both primary keys, with a join on MCU, ITM and LOTN.
This form is mainly used to search by lot number (LOTN), most of the time only with the begining of the LOTN.
The problem is that such search has really poor performance !!
I took a look at it.
When looking at the jdedebug of my fat client, a request is taking more than 30 seconds. If i execute exactly the same request in an sql editor, it takes something like 0,03 seconds !!
With Toad i look at the explain plans. There are totally different for the same request in an sql editor and in jde.
In the sql editor, it uses the index #4 of the F41021 and then the primary key of the F4108. In jde, it process a fast full scan of the index #5 (contains more fields and is less efficient) and after process a full scan of the table F4108 !
Does anyone knows how is working JDE in that case and if it's possible to improve performance by using the good indexes ?
Thanks for your help.
We created a custom interactive application with a form, based on a custom view. This view is based on the F41021 (Item Location File) and the F4108 (Lot Master), including both primary keys, with a join on MCU, ITM and LOTN.
This form is mainly used to search by lot number (LOTN), most of the time only with the begining of the LOTN.
The problem is that such search has really poor performance !!
I took a look at it.
When looking at the jdedebug of my fat client, a request is taking more than 30 seconds. If i execute exactly the same request in an sql editor, it takes something like 0,03 seconds !!
With Toad i look at the explain plans. There are totally different for the same request in an sql editor and in jde.
In the sql editor, it uses the index #4 of the F41021 and then the primary key of the F4108. In jde, it process a fast full scan of the index #5 (contains more fields and is less efficient) and after process a full scan of the table F4108 !
Does anyone knows how is working JDE in that case and if it's possible to improve performance by using the good indexes ?
Thanks for your help.