Wildmatch and ** returning 0

[Mahamed_Qlik]

Hi All,

I am facing high reload time issue using wildmatch function in backend script in below :

if(wildmatch(Attribute, '*DATE*') or wildmatch(Attribute, '*date*') or wildmatch(Attribute, '*Date*'),
date(floor([Attribute Value]),'MM/DD/YYYY'),
if(wildmatch(Attribute, '*prem*') or wildmatch(Attribute, '*PREM*') or wildmatch(Attribute, '*Prem*'),
text(num([Attribute Value],'$#,##0.00')),
text([Attribute Value]))) as [Attribute Value],

It is taking 1hr to reload and if i skip this wildmatch then it taked hardly 10 mins.

Can you please how I can fix this issue and what I can user instead of wildmatch in script

[sunny_talwar]

So, what I am trying to say is that create this new field before you create a qvd.... for example

Table:
LOAD FieldA,
FieldB,
....,
Pick(WildMatch(Upper(Attribute), '*DATE*', '*PREM*') + 1,
  Text([Attribute Value]),
  Date(Floor([Attribute Value]), 'MM/DD/YYYY'),
  Text(Num([Attribute Value], '$#,##0.00'))
) as [Attribute Value]
FROM XYZ;

STORE Table into XYZ.qvd (qvd);

and now load from your qvd with Attribute Value already fixed in it....

[Mahamed_Qlik]

Table:
LOAD FieldA,
FieldB,
....,
Pick(WildMatch(Upper(Attribute), '*DATE*', '*PREM*') + 1,
Text([Attribute Value]),
Date(Floor([Attribute Value]), 'MM/DD/YYYY'),
Text(Num([Attribute Value], '$#,##0.00'))
) as [Attribute Value]
FROM XYZ.qvd;

This is my script.

[sunny_talwar]

Did you try changing it to what I mentioned 🙂

[Mahamed_Qlik]

Yes...It is still taking the same time. 😞

[sunny_talwar]

What is the new script? Can you share both the scripts you use to create the qvd and then the script you use to load the qvd?

[Mahamed_Qlik]

see below is the script.

The comments line is creating issue. Just below it I put yours but that is again giving me same result

LOAD
Key,
factkey,
Attribute,
// if(wildmatch(Attribute, '*DATE*') or wildmatch(Attribute, '*date*') or wildmatch(Attribute, '*Date*'),
// date(floor([Attribute Value]),'MM/DD/YYYY'),
// if(wildmatch(Attribute, '*prem*') or wildmatch(Attribute, '*PREM*') or wildmatch(Attribute, '*Prem*'),
// text(num([Attribute Value],'$#,##0.00')),
// text([Attribute Value]))) as [Attribute Value],
Pick(WildMatch(Upper(Attribute), '*DATE*', '*PREM*') + 1,
Text([Attribute Value]),
Date(Floor([Attribute Value]), 'MM/DD/YYYY'),
Text(Num([Attribute Value], '$#,##0.00'))
) as [Attribute Value],
if(len(num([Attribute Value],'###,###,###,###,###'))>0,[Attribute],null()) as [Attribute Num],
num([Attribute Value],'###,###,###,###,###') as [Attribute Value Num]
FROM
$(vQVDPath)Fact.qvd(qvd);

[sunny_talwar]

Are you having difficulty understand my suggestion? I don't even think you are even trying to do what I have proposed.

[Mahamed_Qlik]

Yes I got it, you mean to say that I should create this new field at the time of QVD store.
But, I am storing this QVD at ETL so dont you think it will good practice to do such transformation in ETL script ?