KeyError when converting df to_list

May 13, 2024

Trying to get one column of a df into a list, but a KeyError is thrown each time.

Writing a function to convert a dataframe to a subplot.
I want to take the first and last value in the df column containing x-axis values so that I can label the xaxis of the subplot with the first and last values only.
tolist() (or to_list) is throwing a KeyError, however.
Can anyone point out where i’m going wrong?

def df_to_subplot(df, df_field_x_axis, subplot_title, nrows, ncols):

    fig, axes = plt.subplots(nrows=nrows, ncols=ncols, sharex=True, sharey=True)
    df.set_index(df_field_x_axis, inplace=True)
    for col, ax in zip(df.columns, axes.flat):
        if col != df_field_x_axis:
            df[col].plot.line(ax=ax)
            ax.set_title(col)

    df = df.sort_values(df_field_x_axis)
    x_ticks = df[df_field_x_axis].tolist()     <<<<<------throws a KeyError
    for ax in axes.flatten():
        ticks = ax.get_xticks()
        ax.set_xticks(x_ticks[0], x_ticks[-1])
        ax.set_yticks()
        plt.setp(ax.get_xticklabels(), rotation=0)

>Solution :

The KeyError is thrown because you’re trying to access a column that doesn’t exist in the DataFrame. This is because you’ve set df_field_x_axis as the index of the DataFrame with df.set_index(df_field_x_axis, inplace=True). After this operation, df_field_x_axis is no longer a column in the DataFrame, but the index.

To fix this, you should get the index values to a list instead of trying to access it as a column. Here’s the corrected code:

def df_to_subplot(df, df_field_x_axis, subplot_title, nrows, ncols):

    fig, axes = plt.subplots(nrows=nrows, ncols=ncols, sharex=True, sharey=True)
    df.set_index(df_field_x_axis, inplace=True)
    for col, ax in zip(df.columns, axes.flat):
        if col != df_field_x_axis:
            df[col].plot.line(ax=ax)
            ax.set_title(col)

    df = df.sort_index()
    x_ticks = df.index.tolist()     # <<<<<------changed this line
    for ax in axes.flatten():
        ticks = ax.get_xticks()
        ax.set_xticks([x_ticks[0], x_ticks[-1]])  # <<<<<------changed this line
        ax.set_yticks()
        plt.setp(ax.get_xticklabels(), rotation=0)

In the line ax.set_xticks([x_ticks[0], x_ticks[-1]]), I’ve changed x_ticks[0], x_ticks[-1] to [x_ticks[0], x_ticks[-1]] because set_xticks expects a list of x locations for the ticks.

https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.DataFrame.set_index.html