7

My initial dataframe df:

                     discharge1  discharge2
datetime                                   
2018-04-25 18:37:00        5862        4427
2018-04-25 21:36:30        6421        4581
2018-04-25 22:13:00        5948        4779
2018-04-26 00:11:30        5703        4314
2018-04-26 02:27:00        4988        3868
2018-04-26 04:28:30        4812        3823
2018-04-26 06:22:30        4347        3672
2018-04-26 10:50:30        3896        3546
2018-04-26 12:04:30        3478        3557
2018-04-26 14:02:30        3625        3598
2018-04-26 15:31:30        3751        3606

What I want to do is to get my dates a list, array, or series where I can iterate over all elements in my list. So that I can use those dates to access rows in another dataframe df_other, and in the end append them to a new dataframe df_new:

for date in date_list():
    df_new = df_new.append(df_other.iloc[df_other.index.get_loc(date)])

which for a date on my list should be run as:

df_new.append(df_other.iloc[df_other.index.get_loc('2018-04-25 18:37:00')])

I tried making a list using df.index but that returns a Datetimeindex where I can only access each date as:

display(df.index[0])
Timestamp('2018-04-25 18:37:00')

where the Timestamp part ruins my .append call.

Also tried df.index.tolist() but that returns a list of: [Timestamp('2018-04-25 18:37:00'), ...]

maciejwww
  • 1,067
  • 1
  • 13
  • 26
Johan R
  • 137
  • 1
  • 1
  • 7
  • 2
    I'm sorry, but I don't understand the problem. Why are these outputs problematic for you? What are you expecting here instead? – cs95 May 12 '18 at 18:48
  • Do you need [df.index.to_pydatetime()](https://stackoverflow.com/q/22825349/2901002) ? – jezrael May 12 '18 at 18:50

1 Answers1

7

Why don't you just iterate over the rows of the dataframe and just use the index values?

create dataframe:

data = [
['2018-04-25 18:37:00',       5862,        4427],
['2018-04-25 21:36:30',       6421,        4581],
['2018-04-25 22:13:00',       5948,        4779],
['2018-04-26 00:11:30',       5703,        4314],
['2018-04-26 02:27:00',       4988,        3868],
['2018-04-26 04:28:30',       4812,        3823],
['2018-04-26 06:22:30',       4347,        3672],
['2018-04-26 10:50:30',       3896,        3546],
['2018-04-26 12:04:30',       3478,        3557],
['2018-04-26 14:02:30',       3625,        3598],
['2018-04-26 15:31:30',       3751,        3606]
]

data = pd.DataFrame(data, columns=['datetime', 'discharge1', 'discharge2'])
data['datetime'] = data['datetime'].apply(pd.to_datetime)
data = data.set_index('datetime')

then iterate over index, and values:

for index, values in data.iterrows():
    print(index)

output:

2018-04-25 18:37:00
2018-04-25 21:36:30
2018-04-25 22:13:00
2018-04-26 00:11:30
...
Troy D
  • 2,093
  • 1
  • 14
  • 28
  • Thank you! For some reason I didnt manage do use iterrows in the correct way. – Johan R May 13 '18 at 13:21
  • @JohanR If this is what you were looking for you can accept the ans by marking the gray button to green so that others searching for a similar query can benefit – Vivek Kalyanarangan May 13 '18 at 14:08
  • @VivekKalyanarangan Ah thank you. New to this site so didn't even know I could do that ^.^ – Johan R May 13 '18 at 14:41
  • What does the ", values" syntax do? When I didn't use this the print(index) would print the whole row of data as opposed to just the date in the index. – Scott Aug 26 '20 at 07:57
  • 1
    @Scott DataFrame.iterrows() returns a tuple that contains two objects, the index value and a Series containing the data. So by using "index, values" you separate out the index and the data. It's the same as if you did "for row in data.iterrows(): index=row[0] values=row[1] – Troy D Aug 27 '20 at 14:52