saddam

• Aug 2nd, 2016

 

use rollup component.

Roshni

• Aug 31st, 2016

 

Could you please explain how rollup can be work for this?

Anil modala

• Sep 18th, 2016

 

In the roll-up by using last and first functions, You can achieve this.

sharmi

• Sep 26th, 2016

 

Use filter by expression

JSP

• Oct 1st, 2016

 

use two dedup
1> dedup keep first
2> keep last
3> key {}

Prakash Poongavanam

• Oct 13th, 2016

 

In case of no indicator how will you read this file? Conditional DML is possible only if there is record_change indicator.

Chaitanya

• Oct 14th, 2016

 

Use Rollup component in expanded mode and select the first and last records.

Karthik

• Oct 19th, 2016

 

Reformat should do this; using o/p index or indexes

laxminarayana

• Feb 28th, 2017

 

How to use Filter By expression?

sandeepkumar

• Mar 17th, 2017

 

Header-Tailer Without Indicator

1.Input file -->Filter by expression (next_in_sequence()>=1) we can collect header record at outport and de select port: We will get the body+data.

2.Connect to Dedupe keep key as null and keep last, Now we will get the count value to outport

3.Body to dupe port

Input--->Filter by expression --->De dupesort--->Outport

krishna

• Mar 20th, 2017

 

emp.dat|12042016
1/A/1000
2/B/2000
3/C/3000
3
for the above file first write input dml as string("
") newline;
and then the data will be stored in input file
and the connect dedup sort port and one out port of it to a file and other out port of it to a dedup sort in port and do the sort using null key and
and then in the keep format place as first and then the first record will goes means header will go and stays in a file and the left content will goes to second dedup sort and in this same go with null key and this time keep parameter as last then last will goes and stay in a file connected to it and rest the data will goes into other file
input ---- dedup dort... 1) op file2
2)dedupsort ..................1)op file2
2)op3

Faijaz

• Mar 28th, 2017

 

Read the file using DML-->string(
) emp_details;
Add filter by expression and put a condition--> length_of(string_split(emp_details,/))==3.
It will filter out the header and trailer

eswara kalyan

• Mar 29th, 2017

 

We can achieve this in multiple ways:

Method-1
Use a dedup with null key and using the option keep first or keep last u can eliminate the first or last record.

Method-2
Using a rollup you can use the function first (in.data) so that you get the first record from the file. Similarly last (in.data) gets last record

SaiGanesh

• Jun 6th, 2017

 

1. Read the records as a single field.
2. Replicate the data, and connect one flow to Leading Records component. Keep the value as 1, so that it outputs only header record.
3. Join output of step 2 to second flow of replicated data. Connect the output to trash, now header alone is eliminated.
4. Now replicate the data of unused port1 of join used in step3.
5. Connect one flow to Rollup, with null key, and leave input must be sorted as it is. Rollup will return the last record.
6. Join output of step 5 to second flow of replicated data from step 4. Connect the output to trash, now trailer alone is eliminated.
7. Unused port1 dat from Join will give only the data records.

jyotiprasad

 Profile Answers by jyotiprasad 

• Aug 22nd, 2017

 

Using rollup component, pass null key, then the select port contain the last record, the op of deselect port used as a lookup use reformat component and use output index here we use lookup count function to achieve the scenario, else use first and last function in rollup

mehak

• Dec 10th, 2018

 

1) Add Sequence to each line.
Input file: your file having input
Output file with sequence added : out1
2) Create file having first rec
Sort
sequence number position equals 1
Input : Ypur input file
Output: file having header record : out2
3) Add trailer record to file
3) Icetool
In : Your input file
Out:out2
Subst from {in} to {out} keep input last
Output : file having header and trailer record:out2
4)Copy non matching records using join keys
Infile1 : input file
Infile2: outfile
Output: file without header and trailer record : out3

Ketan Khot

• Apr 12th, 2019

 

Hi,
We can achive this with only one reformat component.
graph can be -
Input FIle --> Reformat (with output index)--> output files (1-header, 2-data, 3-trailer)
write below code to separate out header/ data/ trailer records in output index
type header_t =
record
string("|") filename;
string("
") date;
end;
type detail_t =
string("/") f1;
string("/") f2;
string("
")f3;
end;

output_index_out :: output_index(in) =
begin
output_index_out :: if(not is_null(reinterpret_as(header_t,in.line))) 0 //if reinterpret_as function return not null that means input record is header record
else if(not is_null(reinterpret_as(detail_t,in.line))) 1 //if reinterpret_as function return not null that means input record is detail record
else 1
end;
note : reinterpret_as function return null vallue when input data is not formated in given dml

Arun

• Dec 3rd, 2019

 

Use scan to take count(). In output select filter count = 0 and max(count()) which could be got by using a global variable in scan to save the max value.

Divya

• Jul 10th, 2023

 

By using a reformat component

Venkat C

• May 5th, 2024

 

Connect the reformat to generate records with single record to read the file and transform the data using below in reformat and connect to target

 /*Reformat operation*/
out::reformat(in)=
begin
out.id :: string_join(vector_sort_dedup_first(string_split_no_empty(read_file
("Fil.dat
"),"
")),"
");
end;