Mercurial > repos > jay > pdaug_merge_dataframes
view PDAUG_Merge_Dataframes/PDAUG_Merge_Dataframes.py @ 8:4056124db7c8 draft
"planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit d396d7ff89705cc0dd626ed32c45a9f4029b1b05"
author | jay |
---|---|
date | Mon, 10 Jan 2022 04:23:30 +0000 |
parents | b5a9f541c6f8 |
children |
line wrap: on
line source
import glob import pandas as pd import sys def MergeData(infiles, add_class_label, class_label, OutPut): data_frame = pd.DataFrame() if add_class_label == 'True' or add_class_label == 'true': for i, file in enumerate(infiles.split(',')): df1 = pd.read_csv(file,sep='\t') df2 = pd.DataFrame(df1.shape[0]*[i], columns=[class_label]) df3 = pd.concat([df1,df2], axis=1) data_frame = pd.concat([data_frame,df3]) final_DF = data_frame.fillna(0) else: for file in infiles.split(','): df1 = pd.read_csv(file,sep='\t') data_frame = pd.concat([data_frame,df1]) final_DF = data_frame.fillna(0) final_DF.to_csv(OutPut, sep="\t", index=False) if __name__=="__main__": import argparse parser = argparse.ArgumentParser() parser.add_argument("-I", "--infiles", required=True, default=None, help=".tsv") parser.add_argument("-L", "--add_class_label", required=False, default=False, help="Path to target tsv file") parser.add_argument("-C", "--class_label", required=False, default='class_label', help="Path to target tsv file") parser.add_argument("-O", "--OutPut", required=False, default='Out.tsv', help="Path to target tsv file") args = parser.parse_args() MergeData(args.infiles, args.add_class_label, args.class_label, args.OutPut)