使用ffill（或其他方法）在pandas数据框中更新列中的多个值

huangapple
go评论109阅读模式
英文:

Updating MULTIPLE values in a column in pandas dataframe using ffill (or other methods)

问题

Here is the translated code portion you provided:

  1. #line 1
  2. scores_full['Event_BOLD_Duration'] = scores_full['Event'].where(scores_full['Event'].isin(['SRNeg', 'SINeg'])).ffill(limit=7).fillna(scores_full['Event']) 
  3. #line 2
  4. scores_full['Event_BOLD_Duration'] = scores_full['Event'].where(scores_full['Event'].isin(['ArrowTask'])).ffill(limit=4).fillna(scores_full['Event']) 
  5. #line 3
  6. scores_full['Event_BOLD_Duration'] = scores_full['Event'].where(scores_full['Event'].isin(['Rating'])).ffill(limit=1).fillna(scores_full['Event'])

Please note that these lines of code are meant to fill NaN values in the 'Event_BOLD_Duration' column based on conditions from the 'Event' column. If you have any further questions or need assistance with this code, feel free to ask.

英文:

I have a dataframe that can be simplified like this:

  1. TR = [17,18,19,20,21,22,23,24,25,26,27,28,29]
  2. Event = ['SRNeg', np.NaN, np.NaN, np.NaN, np.NaN, np.NaN, np.NaN, np.NaN, 'Rating', np.NaN, np.NaN, 'ArrowTask', np.NaN]
  3. df = pd.DataFrame({'Event':Event,'TR':TR})

I want to fill in some NaNs after the event value with that value. Note that not all NaNs are filled. Here is the ideal output:

  1. TR = [17,18,19,20,21,22,23,24,25,26,27,28,29]
  2. Event = ['SRNeg', np.NaN, np.NaN, np.NaN, np.NaN, np.NaN, np.NaN, np.NaN, 'Rating', np.NaN, np.NaN, 'ArrowTask', np.NaN]
  3. Event_BOLD_Duration = ['SRNeg', 'SRNeg', 'SRNeg', 'SRNeg', 'SRNeg', 'SRNeg', 'SRNeg', 'SRNeg', 'Rating', 'Rating', np.NaN, 'ArrowTask', 'ArrowTask']
  4. df = pd.DataFrame({'Event':Event,'Event_BOLD_Duration':Event_BOLD_Duration,'TR':TR})

Here is the code I have so far to complete the above task.

  1. #line 1
  2. scores_full['Event_BOLD_Duration'] = scores_full['Event'].where(scores_full['Event'].isin(['SRNeg', 'SINeg'])).ffill(limit=7).fillna(scores_full['Event']) 
  3. #line 2
  4. scores_full['Event_BOLD_Duration'] = scores_full['Event'].where(scores_full['Event'].isin(['ArrowTask'])).ffill(limit=4).fillna(scores_full['Event']) 
  5. #line 3
  6. scores_full['Event_BOLD_Duration'] = scores_full['Event'].where(scores_full['Event'].isin(['Rating'])).ffill(limit=1).fillna(scores_full['Event'])

However, each line seems to override the previous line's output. How can I fix this issue?

Thank you!

答案1

得分: 1

你需要创建具有通用模式 ".cumsum()" 的各个分组，然后在每个分组上填充值。为了帮助我们，创建一个限制的字典：

  1. limits = {'SRNeg': 7, 'SINeg': 7, 'ArrowTask': 4, 'Rating': 1}
  3. forward_fill = lambda x: x.ffill(limit=limits[x.iloc[0]]) \
  4.                              if x.iloc[0] in limits else x
  6. df['Event_BOLD_Duration'] = (
  7.     df.groupby(df['Event'].notna().cumsum())['Event']
  8.       .transform(forward_fill)
  9. )

输出：

  1. >>> df
  2.         Event  TR Event_BOLD_Duration
  3. 0       SRNeg  17               SRNeg
  4. 1         NaN  18               SRNeg
  5. 2         NaN  19               SRNeg
  6. 3         NaN  20               SRNeg
  7. 4         NaN  21               SRNeg
  8. 5         NaN  22               SRNeg
  9. 6         NaN  23               SRNeg
  10. 7         NaN  24               SRNeg
  11. 8      Rating  25              Rating
  12. 9         NaN  26              Rating
  13. 10        NaN  27                 NaN
  14. 11  ArrowTask  28           ArrowTask
  15. 12        NaN  29           ArrowTask
英文:

You have to create individual group with the common pattern "<condition>.cumsum()" then fill values on each group. To help us, create a dict of limits:

  1. limits = {'SRNeg': 7, 'SINeg': 7, 'ArrowTask': 4, 'Rating': 1}
  3. forward_fill = lambda x: x.ffill(limit=limits[x.iloc[0]]) \
  4.                              if x.iloc[0] in limits else x
  6. df['Event_BOLD_Duration'] = (
  7.     df.groupby(df['Event'].notna().cumsum())['Event']
  8.       .transform(forward_fill)
  9. )

Output:

  1. >>> df
  2.         Event  TR Event_BOLD_Duration
  3. 0       SRNeg  17               SRNeg
  4. 1         NaN  18               SRNeg
  5. 2         NaN  19               SRNeg
  6. 3         NaN  20               SRNeg
  7. 4         NaN  21               SRNeg
  8. 5         NaN  22               SRNeg
  9. 6         NaN  23               SRNeg
  10. 7         NaN  24               SRNeg
  11. 8      Rating  25              Rating
  12. 9         NaN  26              Rating
  13. 10        NaN  27                 NaN
  14. 11  ArrowTask  28           ArrowTask
  15. 12        NaN  29           ArrowTask

答案2

得分: 0

  1. 尝试这样做：
  2. ```python
  3. df['Event_BOLD_Duration']=df['Event'].fillna(method='ffill')
  4. df.loc[10,'Event_BOLD_Duration']=np.nan
  5. print(df)

输出

  1.     Event   TR   Event_BOLD_Duration
  2. 0   SRNeg   17   SRNeg
  3. 1   NaN     18   SRNeg
  4. 2   NaN     19   SRNeg
  5. 3   NaN     20   SRNeg
  6. 4   NaN     21   SRNeg
  7. 5   NaN     22   SRNeg
  8. 6   NaN     23   SRNeg
  9. 7   NaN     24   SRNeg
  10. 8   Rating  25   Rating
  11. 9   NaN     26   Rating
  12. 10  NaN     27   NaN
  13. 11  ArrowTask 28 ArrowTask
  14. 12  NaN     29   ArrowTask
英文:

try this:

  1. df['Event_BOLD_Duration']=df['Event'].fillna(method='ffill')
  2. df.loc[10,'Event_BOLD_Duration']=np.nan
  3. print(df)

Output

  1.     Event	   TR	Event_BOLD_Duration
  2. 0	SRNeg	    17	SRNeg
  3. 1	NaN	        18	SRNeg
  4. 2	NaN	        19	SRNeg
  5. 3	NaN	        20	SRNeg
  6. 4	NaN	        21	SRNeg
  7. 5	NaN	        22	SRNeg
  8. 6	NaN	        23	SRNeg
  9. 7	NaN	        24	SRNeg
  10. 8	Rating	    25	Rating
  11. 9	NaN	        26	Rating
  12. 10	NaN	        27	NaN
  13. 11	ArrowTask	28	ArrowTask
  14. 12	NaN	        29	ArrowTask

huangapple
  • 本文由 发表于 2023年6月29日 21:57:02
  • 转载请务必保留本文链接：https://go.coder-hub.com/76581727.html
  • pandas
  • python
匿名

发表评论

匿名网友

:?: :razz: :sad: :evil: :!: :smile: :oops: :grin: :eek: :shock: :???: :cool: :lol: :mad: :twisted: :roll: :wink: :idea: :arrow: :neutral: :cry: :mrgreen:

确定