@RizkyW08 you just simple create bashscript for split the csv, and call the bashscript from pega. create new function for call the bashscript.

#!/bin/bash
FILENAME=10000.csv
HDR=$(head -1 $FILENAME)   # Pick up CSV header line to apply to each file
split -l 20 $FILENAME xyz  # Split the file into chunks of 20 lines each
n=1
for f in xyz*              # Go through all newly created chunks
do
   echo $HDR > Part${n}    # Write out header to new file called "Part(n)"
   cat $f >> Part${n}      # Add in the 20 lines from the "split" command
   rm $f                   # Remove temporary file
   ((n++))                 # Increment name of output part
done

@RizkyW08 Instead of converting a csv file into multiple files and processing we have option to split csv file data into multiple batches of data and then we can process. Is that design fine for you? Or is it mandatory for you to go with splitting of csv file? What I a, suggesting is splliting of csv data instead of csv file. Please refer the below link for processing large number of records in small batches .

https://docs-previous.pega.com/data-management-and-integration/85/best-practices-processing-files-using-file-listener?