python - 如何忽略日期不在 Pandas 索引中的错误？

Question

我目前有一套不包括圣诞节。我尝试对数据进行子集化，但收到不在索引中的错误。如何设置 pandas 以忽略此错误并将仍在索引中的所有日期子集化？

错误:

KeyError: "DatetimeIndex(['2015-12-25', '2016-12-25'], dtype='datetime64[ns]', freq=None) not in index"

例子:

df = df[pd.date_range(date(2015,6,1), date(2017,8,15))]

Answer 1

你需要:

按 intersection 选择列:

rng = pd.date_range(date(2015,6,1), date(2017,8,15))
df = df[rng.intersection(df.columns)]

或通过exact indexing :

df = df.loc[:, '2015-06-01':'2017-08-15']

或按条件:

df = df.loc[:, (df.columns >= '2015-06-01') & (df.columns <= '2017-08-15')]

---

rng = pd.date_range(datetime(2015,6,1),datetime(2015,6,7))
df = df.iloc[:, df.columns.isin(rng)]

或通过truncate :

df = df.truncate('2015-06-01','2017-08-15', axis=1)

示例:

from datetime import datetime

np.random.seed(452)
rng = pd.date_range('2015-06-01', periods=10)
df = pd.DataFrame(np.random.randint(10, size=(10,10)), columns=rng).iloc[:, np.r_[0:2, 5:9]]
print (df)
   2015-06-01  2015-06-02  2015-06-06  2015-06-07  2015-06-08  2015-06-09
0           0           7           0           3           0           7
1           8           9           8           1           0           2
2           5           2           2           0           0           9
3           2           9           3           8           0           6
4           8           8           7           4           9           8
5           9           9           0           4           0           4
6           2           1           4           1           0           1
7           4           1           9           5           6           7
8           5           9           8           1           4           6
9           6           5           2           5           3           1

---

rng = pd.date_range(datetime(2015,6,1),datetime(2015,6,7))
df1 = df[rng.intersection(df.columns)]

df2 = df.loc[:, '2015-06-01':'2015-06-07']

df3 = df.loc[:, (df.columns >= '2015-06-01') & (df.columns <= '2015-06-07')]

rng = pd.date_range(datetime(2015,6,1),datetime(2015,6,7))
df4 = df.iloc[:, df.columns.isin(rng)]

df5 = df.truncate('2015-06-01','2015-06-07', axis=1)

print (df1)
#print (df2)
#print (df3)
#print (df4)
#print (df5)

   2015-06-01  2015-06-02  2015-06-06  2015-06-07
0           0           7           0           3
1           8           9           8           1
2           5           2           2           0
3           2           9           3           8
4           8           8           7           4
5           9           9           0           4
6           2           1           4           1
7           4           1           9           5
8           5           9           8           1
9           6           5           2           5