[FIXED] Dictionary manipulation, retrieve from list problem

Issue

I have a dictionary linking coordinates to a column name from a dataframe

label_xpos = dict(zip(sensor_label, sensor_x_coordinates))

I want to detect in my df, the empty column, and for each instance, create a list of valid coordinates

for i in range (0,10): 
    nan_columns = df.columns[df.iloc[i].isna()].tolist()
    print(nan_columns)
    xclean = ??

It is silly really because I can do it the other way around (finding the points that are not correct)

xclean = [label_xpos[column] for column in nan_columns]

Thank you

Solution

It is silly really because I can do it the other way around..

If I understand you correctly, you need notna that is the opposite of isna :

for i in range (0,10): 
    notna_columns = df.columns[df.iloc[i].notna()].tolist()
    xclean = [label_xpos[column] for column in notna_columns]

Note that you can avoid processing each row separately and use @ dot product :

from operator import itemgetter

notna_cols = (df.notna() @ (df.columns + ",")).str[:-1].str.split(",").to_dict()
# {0: ['A', 'B', 'C'], 1: ['A', 'C'], 2: ['B', 'C']}

for index, cols in notna_cols.items():
    print(f"{index=}:", itemgetter(*cols)(label_xpos))

# index=0: (11, 22, 33)
# index=1: (11, 33)
# index=2: (22, 33)

Used input :

import pandas as pd

sensor_label = list("ABC")
sensor_x_coordinates = [11, 22, 33]
label_xpos = dict(zip(sensor_label, sensor_x_coordinates))

df = pd.DataFrame([[1, 2, 3], [4, None, 6], [None, 8, 9]], columns=sensor_label)

#      A    B  C
# 0 1.00 2.00  3
# 1 4.00  NaN  6
# 2  NaN 8.00  9

[3 rows x 3 columns]

Answered By - Timeless