Jdbc

[Former Member]

Hi All,

I need help in writing a UDF to append 20 arrays of values.

I want a UDF1 which appends 5 values ,which I pass to second UDF which takes the 5 values from UDF and appends another set of next values.

so it is like cascading effect.

can u help?

Thanks

[justin_santhanam · Accepted solution]

Hi,

This can be done, if you give the structure and the sample data, it could be easy for us to suggest.

Best regards,

raj.

[Former Member]

It will probably do the job well, the only thing that concerns me is that when I run more than one queue UDF in my mappings I run into issues of elements not staying in the right context. I'm going to try and do a little more mapping with this some time, it's been a while since I tried it, so no worries. Thanks Raj.

[Former Member]

Hi Paul,

I am increasing your points since I feel you are following up on this problem.

Thanks

[Former Member]

Hey raj or sap xis, I would like to know how you had the problem solved. I'm just interested in the route taken.

[justin_santhanam]

Paul,

Please find the below URL's that I suggested. Let me know ur concerns,thoguhts regarding this.

http://flickr.com/photo_zoom.gne?id=1062275408&size=o

http://flickr.com/photo_zoom.gne?id=1062275356&size=o

http://flickr.com/photo_zoom.gne?id=1062275340&size=o

Best regards,

raj.

[Former Member]

here's a more efficient algorythm

String final="";

for(int i=1;i<=a.length;i++)

{

final+=a[ i ] ;

if(i%5==0)

{

final+=recorddelimiter;

}

else

{

final+=fielddelimiter;

}

count++;

}

result.addValue(final);

[Former Member]

Raj : Thanks for solving the problem.

Paul: Thanks for being proactively involved.

giving both of you the points.

Thanks again

[Former Member]

Hmmm, that's probably true... I hope Raj has something for you then.

It'd seem to me though, that running it through two UDFs would make it more congested from a speed standpoint especially since you'll be using Advanced UDFs both times pulling in the contexts.

[Former Member]

I still think the beast way would be to do it with one UDF so you don't get the queuing issues.

Sourcenode -> removeContexts-> UDFCombine -

---

> DestinationNode

in the node use this code

String final="";

int grab=0;

while(grab<a.length)

{

for(int j=0;j<5;j++)

{

final+=a[grab];

if(j<4)

{

final+=fielddelimiter;

}

grab++;

}

final+=recorddelimiter;

}

result.addValue(final);

[Former Member]

Issue is in passing a large number of values to a user defined function.

before passing i am mapping accoriding to a logic.

so it looks so much congested.

Thanks

[Former Member]

If you have a document structure like this

Root

--next level (5 structures of this)

-

---

next level (20 elements of this)

I have an easy solution.

[Former Member]

Hi ,

source is a material master IDOC.

target is a stored procdure.

I have done all the mappings.

But now stored procdure needs all the values appended together and sent as a string with a record delimiter of | and field delimiter of ,

I dont want one append function since there is a complex mapping for each field. so I cannot pass all mapped values to Append udf in one go.

For this reason I need 2-3 smilar UDFs : first udf could append 5 values and this culd be passed to second udf which appends the result from first udf to second set of values.

Thanks

[Former Member]

The mapping was working fine when I was making calls to store procedure before.

but for each cross reference segment(repeating) a call had to be made to stored procedure.

this led to an overhead of connecting to stored procedure many times.

but now I have to append all together in one string and only one call to stored procedure.

Ask me more questions if you have more doubts

[Former Member]

I'd like to see the structure of these documents. I can probably help you if I understand this better.