IDI_Master/projet_INSEE
1
0
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity

Process towns DataFrame

Browse source 
Add area colum from statistics
This commit is contained in:
Yorick Barbanneau 2022-04-19 22:40:35 +02:00
parent 77456db238
commit e0b69e18f1
1 changed files with 17 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

20
create_db.py
View file

@ -111,6 +111,7 @@ def import_statistics_csv(raw_file):
""" """
logger.info('import town from {}'.format(raw_file)) logger.info('import town from {}'.format(raw_file))
stats_convert= lambda x: x if len(str(x)) == 5 else f'0{x}'
stats = pd.read_csv(raw_file, stats = pd.read_csv(raw_file,
usecols=["CODGEO","SUPERF","P18_POP","P13_POP","P08_POP","D99_POP", usecols=["CODGEO","SUPERF","P18_POP","P13_POP","P08_POP","D99_POP",
"NAIS1318","NAIS0813","NAIS9908","NAIS9099","NAIS8290","DECE1318", "NAIS1318","NAIS0813","NAIS9908","NAIS9099","NAIS8290","DECE1318",
@ -120,7 +121,9 @@ def import_statistics_csv(raw_file):
"P13_RP","P08_RP","D99_RP","D90_RP","D82_RP", "P18_RSECOCC", "P13_RP","P08_RP","D99_RP","D90_RP","D82_RP", "P18_RSECOCC",
"P13_RSECOCC","P08_RSECOCC","D99_RSECOCC","D90_RSECOCC", "P13_RSECOCC","P08_RSECOCC","D99_RSECOCC","D90_RSECOCC",
"D82_RSECOCC"], "D82_RSECOCC"],
sep=';') sep=';',
converters={'CODGEO':stats_convert}
)
return stats return stats
def get_single_date(attr): def get_single_date(attr):
@ -241,8 +244,7 @@ if __name__ == '__main__':
## create statistics dataframes ## create statistics dataframes
# #
# We need to first iterate on statistics # We need to first iterate on statistics
if args.verbose or args.debug: t.start('Process_Statistics')
t.start()
c_stats = pd.DataFrame(columns = ['com','id_indicateur','date_debut', c_stats = pd.DataFrame(columns = ['com','id_indicateur','date_debut',
'date_fin','valeur'] 'date_fin','valeur']
@ -290,6 +292,18 @@ if __name__ == '__main__':
temp['valeur'].append(value) temp['valeur'].append(value)
t.stop() t.stop()
t.start('Process_town')
print(statistics[['SUPERF','CODGEO']])
towns = pd.merge(towns,
statistics[['CODGEO', 'SUPERF']],
left_on=['COM'],
right_on=['CODGEO'],
how = 'left'
)[['COM','NCC','LIBELLE', 'DEP', 'SUPERF']]
t.stop()
logger.debug(towns)
t.get_total_time() t.get_total_time()
sys.exit() sys.exit()