迭代地更改Pandas数据框的列中的每个单元格

Question

I am trying to change the value of every cell in a Pandas data frame. Expecting .loc to allow me to identify a cell with the paradigm df.loc[row_index, column_name] = cell value , I've used the following loop:

table["field"] = 6 #placehodler value used only to create the column
for field, index in enumerate(table["field"]):   table.loc[index,
"field"] = table.loc[index, "field_x"] if math.isnan(table.loc[index,
"field_y"]) else table.loc[index, "field_y"]

However, I'm getting the following error: KeyError: 'the label [6] is not in the [index]' . What is the correct syntax for selecting values by index?

Answer 1

You shouldn't be using enumerate to generate the index and column values, you should use iterrows .

Example usage:

In [6]:
df = pd.DataFrame({'a':np.arange(5), 'b':np.random.randn(5)})
df

Out[6]:
   a         b
0  0 -0.579585
1  1 -0.582196
2  2 -0.367147
3  3 -0.363332
4  4  0.880826

In [9]:
for index, row in df.iterrows():
    print('index: ', index)
    print('row: ', row)

output:

index:  0
row:  a    0.000000
b   -0.579585
Name: 0, dtype: float64
index:  1
row:  a    1.000000
b   -0.582196
Name: 1, dtype: float64
index:  2
row:  a    2.000000
b   -0.367147
Name: 2, dtype: float64
index:  3
row:  a    3.000000
b   -0.363332
Name: 3, dtype: float64
index:  4
row:  a    4.000000
b    0.880826
Name: 4, dtype: float64

This will allow you to access a specifc row using df.loc[index] , if you need the columns you can access using row.index in the above for loop