subhasis nag

• Feb 10th, 2017

 

using first_defined function. it will take out only not null values

subhasis nag

• Feb 13th, 2017

 

Using FBE giving the condition !in_null() will pass the not null value only

Rishiraj

• Feb 13th, 2017

 

With rollup select key as {Id} and in transform do following:
string_filter_out(concatenation(in.Col1),"NULL").
Same for Col2.

Naveen

• Feb 22nd, 2017

 

Hi Rishiraj,

The code which posted will not work......
However the approach is correct to achieve the output rather than using the string_filter_out we need to use string_filter like below....

/*Do computation*/
out::rollup(in)=
begin
out.* :: in.*;
out.col1 :: string_filter(concatenation(in.col1),0123456789);
out.col2 :: string_filter(concatenation(in.col2),0123456789);
end;

swapnil5291

 Profile Answers by swapnil5291 

• Feb 23rd, 2017

 

We can use rollup component here.
ID as key for rollup
And simply use concatenation function for col1 and col2.

Anirban

• May 31st, 2017

 

Please use the below code:
I/p file -> rolup (in rollup we can use accumulation to create vector) -> reformat (use vector_sort_dedup_first) -> normalize

geet

• Jul 29th, 2017

 

Use roll up with Id as the key, and roll up function as below,
out.col1:: concatenation(in.col1, !is_null(in.col1));
out.col2 :: concatenation(in.col2, !is_null(in.col2)):

Bhanu

• Nov 7th, 2017

 

/*Do computation*/
out::rollup(in)=
begin
out.* :: in.*;
out.amount:: concatenation(string_filter(in.amount,"0123456789"));
out.f:: concatenation(string_filter(in.f,"0123456789"));
end;

Sanjay Botcha

• Nov 26th, 2018

 

input --reformat -- rollup -- output
Reformat :
out.id = in.id;
out.col1 :: if(is_null(in.col1))0 else in.col1;
out.col2 :: if(is_null(in.col2))0 else in.col2;
Rollup:
Summarize on key as id

M Farhan

• Dec 29th, 2018

 

Sort. By Id col1 col2
Then dedup keep last !

Srushti Makhija

• Jun 13th, 2019

 

Use first function in rollup

Amaan Ansari Exusia

• Aug 21st, 2019

 

Input-> Sort (key : id) -> Roll up (ID KEY) (Concat : col1 concat col2) -> Output

Shubham Gautam

• Sep 16th, 2019

 

use rollup on Id column>>> in the rollup transform do-
out.Id::in.Id
out.col1::sum(in.col1)
out.col2::sum(in.col2)

Shri

• Feb 28th, 2020

 

out :: rollup(in) =
begin
out.name :: in.name;
out.marks1::max(in.marks1);
out.marks2::max(in.marks2);
end;

Bibhu Prasad Das

• Mar 27th, 2020

 

Easiest way is to replace the "Null" to a finite value i.e zero preferably and then rollup the data based on id as key to find max value of the fields to get the result easily. Please read whole file in string instead of using decimal datatype for second and third field.
Flow :-
input --> reformat --> sort --> rollup --> output
Reformat : -> replace null with Zero
out.id = in.id;
out.col1 :: if((in.col1=="Null")"0" else in.col1;
out.col2 :: if((in.col2=="Null")"0" else in.col2;
sort : -> id
Rollup: -> use id as key
/*Summarize on key as id*/
/*Do computation*/
out::rollup(in)=
begin
out.* :: in.*;
out.col1 :: max(in.col1);
out.col2 :: max(in.col2);
end;
P.S - You cant operate on NULL in rollup unless you handle them prior to rollup. Here its only a string "Null" and not NULL value. So please give substantial thought on coming up with answer.
Additionally i have tried above stuff and it gave me proper result.

Shikha

• Jun 23rd, 2020

 

Use Input file --> Sort(key - id) --> Rollup (key -id) --> Output file
In Rollup :-
out :: rollup(in)=
begin
out.id :: in.id;
out.col1 :: concatenation(in.col1);
out.col2:: concatenation(in.col2);
end;

Sonal Singh

• Aug 8th, 2020

 

You can directly FBX and add !is_null(field1) or !is_null(field2)

Divya

• Oct 27th, 2021

 

Use rollup
Key : {id}
transform :
type temporary_type =
record
string("|") prev_record_col1;
string("|") prev_record_col2;
end;
temp :: initialize(in) =
begin
temp.prev_record_col1 :: allocate_with_defaults();
temp.prev_record_col2 :: allocate_with_defaults();
end;
out :: rollup(temp, in) =
begin
out.prev_record_col1 :: if(!is_null(in.col1)) string_concat(temp.prev_record_col1,in.col1) else temp.prev_record_col1;
out.prev_record_col2 :: if(!is_null(in.col2)) string_concat(temp.prev_record_col2,in.col2) else temp.prev_record_col2;
end;
out :: finalize(temp, in) =
begin
out.id :: in.id;
out.col1 :: temp.prev_record_col1;
out.col2 :: temp.prev_record_col2;
end;

Swapnil

• Feb 9th, 2023

 

One component is enough for this - Rollup
Rollup Key - {id}
in transform:
out.id :: in.id;
out.col1:: sum(first_defined(in.col1,0));
out.col2:: sum(first_defined(in.col2,0));
end;

reddy,sudhakar

• Mar 28th, 2023

 

Divya


Use rollup
Key : {id}
transform :
type temporary_type =
record
string("|") prev_record_col1;
string("|") prev_record_col2;
end;
temp :: initialize(in) =
begin
temp.prev_record_col1 :: allocate_with_defaults();
temp.prev_record_col2 :: allocate_with_defaults();
end;
out :: rollup(temp, in) =
begin
out.prev_record_col1 :: if(!is_null(in.col1)) string_concat(temp.prev_record_col1,in.col1) else temp.prev_record_col1;
out.prev_record_col2 :: if(!is_null(in.col2)) string_concat(temp.prev_record_col2,in.col2) else temp.prev_record_col2;
end;
out :: finalize(temp, in) =
begin
out.id :: in.id;
out.col1 :: temp.prev_record_col1;
out.col2 :: temp.prev_record_col2;
end;

Vishnu

• Dec 12th, 2023

 

Does it works on GDE 1.1

Iffath

• Dec 13th, 2023

 

Use Rollup with key as id and in transform use sum() function.
mention fields as nullable in dml

Code
out::rollup(IN) =


BEGIN


out.id :: IN.id;


out.col1 :: sum(IN.col1);


out.col2 :: sum(IN.col2);


END;


 


USE below dml 


record 


string(|)id;


DECIMAL(|)col1 = NULL("");


DECIMAL(


)col2 = NULL("");


END;