Pandas Drop all rows after all rows == True-CodePudding

I have a df like this:

        Low isLower  isPrevHigher  isNextHigher
0   22470.0    True         False          True
1   22480.0     NaN         False          True
2   22576.6     NaN         False          True
3   22600.4     NaN         False         False
4   22583.5     NaN          True          True
5   22652.2     NaN         False          True
6   22656.8     NaN         False         False
7   22646.5     NaN          True         False
8   22600.0     NaN          True         False
9   22555.0     NaN          True          True
10  22580.1     NaN         False          True
11  22620.0     NaN         False          True
12  22682.2     NaN         False         False
13  22681.0     NaN          True          True
14  22710.8     NaN         False         False
15  22657.2     NaN          True         False
16  22623.0     NaN          True          True
17  22634.0     NaN         False          True
18  22660.0     NaN         False          True
19  22673.6     NaN         False          True
20  22721.2     NaN         False         False
21  22580.0     NaN          True         False
22  22552.6     NaN          True         False
23  22382.6    True          True         False
24  22353.0    True          True         False
25  22341.7    True          True         False
26  22312.4    True          True         False
**27  22256.4    True          True          True**
28  22310.6    True         False         False
29  22286.0    True          True          True
30  22306.8    True         False          True
31  22386.3    True         False         False

I want to drop all rows after the first isLower == True & isPrevHigher == True & isNextHigher == True.

So everything after row 27.

CodePudding user response：

drop_row = df[df[['isLower', 'isPrevHigher', 'isNextHigher']].eq(True).all(axis=1)].index[0]
df = df[df.index <= drop_row]
print(df)

Output:

        Low isLower  isPrevHigher  isNextHigher
0   22470.0    True         False          True
1   22480.0     NaN         False          True
2   22576.6     NaN         False          True
3   22600.4     NaN         False         False
4   22583.5     NaN          True          True
5   22652.2     NaN         False          True
6   22656.8     NaN         False         False
7   22646.5     NaN          True         False
8   22600.0     NaN          True         False
9   22555.0     NaN          True          True
10  22580.1     NaN         False          True
11  22620.0     NaN         False          True
12  22682.2     NaN         False         False
13  22681.0     NaN          True          True
14  22710.8     NaN         False         False
15  22657.2     NaN          True         False
16  22623.0     NaN          True          True
17  22634.0     NaN         False          True
18  22660.0     NaN         False          True
19  22673.6     NaN         False          True
20  22721.2     NaN         False         False
21  22580.0     NaN          True         False
22  22552.6     NaN          True         False
23  22382.6    True          True         False
24  22353.0    True          True         False
25  22341.7    True          True         False
26  22312.4    True          True         False
27  22256.4    True          True          True

CodePudding user response：

You may want to drop rows on/after the first row with ALL empty values:

    # create another data frame
df = pd.DataFrame(
    {'direction': ['north', 'east', 'south', None, 'up', 'down'],
     'amount': [10, 20, 30, None, 100, 200]})

# does the whole row consist of `None`
df['row_is_none'] = df.isna().all(axis=1)

# calculate the cumulative sum of the new column
df['row_is_non_accum'] = df['row_is_none'].cumsum()

# create boolean mask and perform drop (not shown to save space)

print(df)
  direction  amount  row_is_none  row_is_non_accum
0     north    10.0        False                 0
1      east    20.0        False                 0
2     south    30.0        False                 0
3      None     NaN         True                 1
4        up   100.0        False                 1
5      down   200.0        False                 1

CodePudding user response：

This will find the where all the specified columns = True, then find the lowest index making it into a number variable the using iloc to final all the index's specified

first_all_true = df.iloc[np.where((df['isLower'] == True) & (df['isPrevHigher'] == True) & (df['isNextHigher'] == True))]index[0]
df.iloc[0:first_all_true   1]

CodePudding user response：

Using boolean indexing with help of all, the boolean NOT (~), and cummin:

df[(~df[['isLower', 'isPrevHigher', 'isNextHigher']].eq(True).all(1)).cummin()]

NB. untested answer