Pandas: Is there a way to iterate through a dataframe and create new dataframes using multiple condi-CodePudding

I need to split a dataframe up so that I can create new files using the spatial data inside the full dataframe.

I have a dataframe that looks like this:

Full dataframe

I'd like to create some new dataframes using multiple conditions

I'm hoping to end up with a bunch of dataframes with conditions where, for example:

Z_year = 2020, Z_month = 5, Z_group = 'weekday', time_range = '08:00 to 12:00'

Z_year = 2020, Z_month = 5, Z_group = 'weekday', time_range = '12:00 to 18:00'

Z_year = 2020, Z_month = 5, Z_group = 'weekday', time_range = '18:00 to 01:00'

Z_year = 2020, Z_month = 5, Z_group = 'weekend', time_range = '08:00 to 12:00'

Z_year = 2020, Z_month = 5, Z_group = 'weekend', time_range = '12:00 to 18:00'

Z_year = 2020, Z_month = 5, Z_group = 'weekend', time_range = '18:00 to 01:00'

Z_year = 2020, Z_month = 8, Z_group = 'weekday', time_range = '08:00 to 12:00'

Z_year = 2020, Z_month = 8, Z_group = 'weekday', time_range = '12:00 to 18:00'

Z_year = 2020, Z_month = 8, Z_group = 'weekday', time_range = '18:00 to 01:00'

...

Z_year = 2021, Z_month = 11, Z_Group = 'weekend', time_range = '18:00 to 01:00'

etc.

Until all combinations are complete.

Is there a way to iterate through my dataframe to do this?

The target collumns are (Year, month, group, and time range), and the unique values for each are: (2020, 2021), (5, 8, 11), ('weekday', 'weekend'), and (08:00 to 12:00, 12:00 to 18:00, 18:00 to 01:00)

...

#This is used to find unique values for target collumns
year_list = df['Z_year'].unique()

#Creating a new dataframe using only target values
a = df.loc[(df['Z_year'] == 2020) & (df['Z_month'] == 5) & (df.loc['Z_group'] == 'weekday') & (df.loc['time_range'] == '08:00 to 12:00')]

...

Thanks for your help!

CodePudding user response：

You can iterate through the dataframe like this:

cols, rows = df.shape
for x in range(cols):
    print(df[x:x 1])

You can select and append candidates into a list and convert it into a new dataframe.

CodePudding user response：

I suggest to use a for loop and a dictionary so to save each dataframe with a clef composed of the combination of the conditions

dic = {}
for a in [2020,2021]:
    for b in [5, 8, 11]:
        for c in ['weekday', 'weekend']:
            for d in ['08:00 to 12:00', '12:00 to 18:00', '18:00 to 01:00']:
                idf = str(a) str(b) c d
                dic[idf] = df.loc[(df['Z_year'] == a) & (df['Z_month'] == b) & (df.loc['Z_group'] == c) & (df.loc['time_range'] == d)]