Last Updated: 2024-11-23

This tip is a simple intuition (that seems near obvious in retrospect) about guessing the argument form in some dependency's API.

I struggled using the Python pandas' .iloc method. My goal was to get the data for a certain column (real_prob was its title) available on a bunch of rows with known index locations.

My code was:

row_indexes = [899, 1044, 482...]
df.iloc[row_indexes, "real_prob"] = 0.9

This failed. The errors didn't help me solve it, but I eventually realized that iloc expected numerical indexes for every dimension. Therefore this would have worked (assuming real_prob was the 4th column - 3rd with 0-indexing)

row_indexes = [899, 1044, 482...]
df.iloc[row_indexes, 3] = 0.9

Or I could have calculated it dynamically:

df.iloc[row_indexes, df.columns.get_loc("real_prob")] = 0.9