java udf

[Former Member]

Hello sdners

I need help with a java udf

context udf : array of texts will be passed to udf

text1 : This is a purchase order number 111111....# and date is 10/02

text2: purchase order number 222..# and date is 10/03

text n...

In all these texts being passed to the udf,we need to extract the string : purchase order 11111...

so basically the substring needs to be extracted "purchase order 11111... " . Anything between purchase order... and #: the substring needs to be extracted.

The substring extracted is dynamic in length : purchase order ..... : can be of any length.

regards

[Former Member · Accepted solution]

Hi,

input text is : The purchase order number is 123456#

output : purchase order number is 123456

if the above is ur requirement then check with the below udf..it works!!

use cache parameter as value

public String find(String a,Container container)
{
   //write your code here
 String start_pos = "purchase order";
String end_pos = "#";
int s1 = a.indexOf(start_pos);
int s2 = a.indexOf(end_pos);
String output = a.substring(s1,s2);
return output;
}

srctext-->find->tgt

[Former Member · Accepted solution]

Hi Malini/Sridhar

Thanks to both of you.

Malini has understood the requirement correctly as her understanding of the question is correct although i have not tried her udf.

But Malini ,the field is repeating : I mean there can be many texts coming as the segment under which it comes is a repeating segment. You gave cache parameter as value ,but I guess since the field is repeating we need to give cache parameter as Context. Can you modify your functions as a context function instead of value ?

Regards

[Former Member · Accepted solution]

Hi Denny

Try this

public void returnPurchaseOrder(String[] input,ResultList result,Container container)
{
    //write your code here
String output="";

  for(int i=0;i<input.length;i++){
      int start = input.indexOf('Purchase order');
      int end = input.indexOf('#');
      output = text1.substring(start,end);
  }
result.addValue(output);
} 

Thanks

Gaurav

[Former Member · Accepted solution]

A general question : When in java we say index of ( purchase order) what is the value returned : postion of p or purchase order?

Regards

[Former Member · Accepted solution]

HI Denny

Thanks for asking this i made a mistake in the code above. Corrected one is

public void returnPurchaseOrder(String[] input,ResultList result,Container container)
{
    //write your code here
String output="";
 
  for(int i=0;i<input.length;i++){
      int start = input.indexOf("Purchase order");
      int end = input.indexOf('#');
      output = text1.substring(start,end);
  }
result.addValue(output);
} 

In Java when we say index of - it return position(integer) of a character or string.

indexOf( ' character' )

indexOf ( "String" )

In case source come as PURCHASE ORDER and sometime purchase Order we can use compareToIgnoreCase("Purchase order") that should solve the problem

Thanks

Gaurav

[Former Member · Accepted solution]

So in this string : " This is a purchase order number 11111 #" what is the postion of this string : " purchase order "

Secondly how to incorporate the logic for ignoring the case of " purchase order" . It is free form text and it can come as PURCHASE ORDER.

Thank You

[Former Member · Accepted solution]

Hi Denny

 This is a purchase order number 11111 # 

in this "purchase order" is 11 but for your requirement to get only 1111 out of the above string you need to keep "purchase order number" into the code i provided above instead of "purchase order"

Secondly for your requirement of the above string you can incorporate

 int start = input.toLowerCase().indexOf("purchase order"); 

Thanks

Gaurav

[Former Member · Accepted solution]

Dear Mr Bhargav

Thank you .

There is an old saying : "That it ends where it starts". You had initally started answering this query

But Malini and Sridhar's inputs were also very significant.

Will award the points now.

Regards

[Former Member]

Hi,

I have modified my udf.

Hope it works 4r repeating segments.

Use cache parameter as context

public void find(String[] a,ResultList result,Container container)
{
 //write your code here
 String start_pos = "purchase order";
String end_pos = "#";
for(int i=0 ; i != a.length ; i++)
{
int s1 = a<i>.indexOf(start_pos);
int s2 = a<i>.indexOf(end_pos);
String output = a<i>.substring(s1,s2);
result.addValue(output);
}
}

[Former Member]

Thank You Mr Balasuamania.

[Former Member]

Hi Denny

thanks for the clarification

you can try this

String out="";

for(int i=0;i<text1.length;i++){

int x = text1<i>.indexOf('#');

out = text1<i>.substring(0,x);

}

result.addValue(out);

Thanks

Gaurav

[Former Member]

Dear Mr Bhargava

In your substring function you have given as (0,x )as parameters ,although "purchase" can come anywhere in the text and not at position 0 only.

Could you clarify ?

Regards

[Former Member]

Hi

Ohh can you give me the some sample data so that we can think of dynamic value for this.

How can we identify the start of purchase order number. like you have specified end as #

Thanks

Gaurav

[Former Member]

Hi

I am not sure but cant we identify in a text : a substring starting with purchase and ending with #

Purchase(wild card *) ends with #

Text1 : This is a purchase order number 111111#

Text2: Purchase order number 2222#

Text3 : This is an invoice but its purchase order number is 33333# and its invoice number is 44.

If we have texts like these passed to your udf ,then can we identify this substring ( Purchase order number ..) based on the logic : starts with purchase and everything before #?

I am not sure though.

Regards

[Former Member]

Hi

Without the info of identifying we need to consider every case out here.

The above UDF code can be enhanced using any other string operation

http://java.sun.com/j2se/1.5.0/docs/api/java/lang/String.html

Thanks

Gaurav

[Former Member]

Does someone have any solution for this unique issue?

Regards

Edited by: denny morris on Oct 4, 2008 3:47 PM

[Former Member]

Hi Denny,

Based on my understanding the output should be Purchase order should return the UDF .

If my understanding is right then the following udf return the Puchase order number.

import java.util.*;

String s="";

StringTokenizer parser = new StringTokenizer(a);

while (parser.hasMoreTokens()) {

String s1=parser.nextToken();

if(s1.endsWith("#")){

s = s1;

}

}

int k = s.length()-1;

return s.substring(0,k);

This will return the Puchase order number in a given any String ends with #.

for the above texts it will give the output as 111111

Regards

Sridhar Goli

[Former Member]

Sridhar:

In your understanding of requirement :

Input : This is a purchase order number 11111 and invoice is 222

Outut : purchase order number 11111

Since this is a free form text ,it needs to be returned in target as : purchase order number 11111

and it needs to be a context function as it can repeat many time with different values as it comes under a repeating segment.

Regards

[Former Member]

Hi Denny

can you provide more info on your requirement

In case you don't have context type UDF this is pretty easy to do

text1 = "11111#222"

for this x = text1.IndexOf('#');

String out ="";

out = text1.substring(0,x);

return out;

Thanks

Gaurav

.

[Former Member]

Dear Mr Bhargava

In source the text field repeats. So i heard ,we need a contxt or queue udf for this since we need to pass a list of values.

Could you specify what else is not clear in the requirement? so that i could explain more.

Regards

[Former Member]

We will have list of texts which i have mentioned and we need to extract only the substring which tells the purchase order number and to specify the end of this substring there is a # coming ,so we need to extract the sub string : purchase order number......

Everything after purchase and before # needs to be extracted.