英文:
How to update group in df after iteration
问题
我试图对每个数据框组应用一个操作，然后更新原始数据帧中的相应行。然而，新值未插入到正确的位置。实际上，我正试图在每个组内进行diff操作。
mydf = pd.read_csv("my.csv")
mydf["forwardDelta"] = np.NaN
for name, group in mydf.groupby(["a", "b"]):
group["forwardDelta"] = group["c"] - group["c"].shift(1)
for index, row in group.iterrows():
mydf.iloc[index, 4] = row["forwardDelta"]
英文:
I am trying to apply an operation to each data frame group, and then update the corresponding rows in the original data frame. However, the new values are not being inserted at the correct locations. Effectively, I am trying to diff within each group.
mydf = pd.read_csv("my.csv")
mydf["forwardDelta"] = np.NaN
for name, group in mydf.groupby(["a", "b"]):
group["forwardDelta"] = group["c"] - group["c"].shift(1)
for index, row in group.iterrows():
mydf.iloc[index, 4] = row["forwardDelta"]
答案1
得分: 1
我认为这里不需要循环 - 使用 DataFrameGroupBy.diff，如果需要设置第5列的值，请使用 DataFrame.iloc 并使用 : 选择所有行：
mydf = pd.DataFrame({'a':[5,8]*3,
'b':[1,2]*3,
'c':[2,7,5,4,3,9],
'd':list('abcdef'),
'e':range(5,11)}).sort_values(['a','b'], ignore_index=True)
print (mydf)
a b c d e
0 5 1 2 a 5
1 5 1 5 c 7
2 5 1 3 e 9
3 8 2 7 b 6
4 8 2 4 d 8
5 8 2 9 f 10
mydf["forwardDelta"] = mydf.groupby(["a", "b"])["c"].diff()
mydf.iloc[:, 4] = mydf["forwardDelta"]
print (mydf)
a b c d e forwardDelta
0 5 1 2 a NaN NaN
1 5 1 5 c 3.0 3.0
2 5 1 3 e -2.0 -2.0
3 8 2 7 b NaN NaN
4 8 2 4 d -3.0 -3.0
5 8 2 9 f 5.0 5.0
英文:
I think here no loops are necessary - use DataFrameGroupBy.diff and if need set values in 5th column use DataFrame.iloc with : for select all rows:
mydf = pd.DataFrame({'a':[5,8]*3,
'b':[1,2]*3,
'c':[2,7,5,4,3,9],
'd':list('abcdef'),
'e':range(5,11)}).sort_values(['a','b'], ignore_index=True)
print (mydf)
a b c d e
0 5 1 2 a 5
1 5 1 5 c 7
2 5 1 3 e 9
3 8 2 7 b 6
4 8 2 4 d 8
5 8 2 9 f 10
mydf["forwardDelta"] = mydf.groupby(["a", "b"])["c"].diff()
mydf.iloc[:, 4] = mydf["forwardDelta"]
print (mydf)
a b c d e forwardDelta
0 5 1 2 a NaN NaN
1 5 1 5 c 3.0 3.0
2 5 1 3 e -2.0 -2.0
3 8 2 7 b NaN NaN
4 8 2 4 d -3.0 -3.0
5 8 2 9 f 5.0 5.0
通过集体智慧和协作来改善编程学习和解决问题的方式。致力于成为全球开发者共同参与的知识库，让每个人都能够通过互相帮助和分享经验来进步。
评论