Convert mm:ss string to hh:mm:ss time format in Python-CodePudding

I want to convert the values in df['Chip Time'] column which are string obj to timedelta64.

Have tried this....> df2['Chip Time'] = pd.to_timedelta(df2['Chip Time'])

BUT GETS Error message ...> ValueError: expected hh:mm:ss format

CodePudding user response：

You can add :00 if string has only one : by Series.str.count with Series.mask:

df2 = pd.DataFrame({"Chip Time": ['10:04','1:45:23','23:12']})

s = df2['Chip Time'].mask(df2['Chip Time'].str.count(':').eq(1), df2['Chip Time']   ':00')
#alternative:
#s = df2['Chip Time']   np.where(df2['Chip Time'].str.count(':').eq(1), ':00', '')
print (s)
0    10:04:00
1     1:45:23
2    23:12:00
Name: Chip Time, dtype: object

df2['Chip Time'] = pd.to_timedelta(s)
print (df2)
        Chip Time
0 0 days 10:04:00
1 0 days 01:45:23
2 0 days 23:12:00

CodePudding user response：

You can use a regex with str.replace to add '0:' on values matching 'x:x' only:

df2['Chip Time'] = pd.to_timedelta(df2['Chip Time'].str.replace('^(\d :\d )$', r'0:\1', regex=True))

Example output:

        Chip Time
0 0 days 00:16:48
1 0 days 01:07:51

Used input:

  Chip Time
0     16:48
1   1:07:51