Output Phase reading - 1 more tip

[Ken Langendock]

I forgot to mention one other trick I learned over the years. OUTPUT PHASE READING AT FINAL.

Let’s say you are creating a QTP extract that has multiple parts/requests. After each request you end up with more fields added to the subfile read in until you get to the final subfile you want.

In this case you would do something like this (Simple view NOT A REAL WORLD EXAMPLE): 

REQUEST 1

ACCESS Orders 

; do something

SUBFILE sf001 include Orders

REQUEST 2

ACCESS *sf001 link to OrderDetails

SUBFILE sf002 include sf001, fields from OrderdDetails

REQUEST 3

ACCESS *sf002 link to Products

SUBFILE sf003 include sf002, fields from Products

 

This method of coding has performance issues. Every time you read a file and output another one, there is more activity with the O/S, Buffers, etc.

BUT, if you take an OUTPUT PHASE READ approach and put all the fields in the first subfile and then each request just updates the fields it needs to you get better performance again (Marginal, but I have proven this back in the 80’s):

REQUEST 1

ACCESS Orders 

; the following two statements open the files with nothing in it at the beginning of the output phase ONCE!!!!.

OUTPUT OrderDetails ADD UPDATE NOITEMS AT FINAL

OUTPUT Products ADD UPDATE NOITEMS AT FINAL

SUBFILE sf001 include Orders, fields from OrderDetails, fields from Products

REQUEST 2

ACCESS *sf001 link to OrderDetails

OUTPUT sf001 

  ITEM fields OF sf001 FINAL fields of OrderdDetails

REQUEST 3

ACCESS *sf001 link to Products

OUTPUT sf001 

  ITEM fields OF sf001 FINAL fields of Products

 

Some people will say “I can just create TEMPS to do that”, but if you are reporting the contents of this subfile in QUIZ, you have to add all the attributes to the columns in your REPORT and FOOTING statements. This way all the attributes come with the subfile. I try to NEVER create a TEMPORARY item that will be included in a subfile if I can.

Again, you know your data and your programs, this just makes the extracts run faster.

Ken

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.sowder.com/pipermail/powerh-l/attachments/20090819/0d614308/attachment.htm