← All lessons

loadtxt and genfromtxt

Upload sample files, delimiters, comments, usecols, and missing tokens.

Download the sample files from the banner, then click Add files in the workbench. Comment lines starting with # are skipped. There are no string city columns — row order is Nairobi, Mombasa, Kisumu, Nakuru, Eldoret.

Goal

List /uploads, load units.csv and temps.txt, and read missing tokens with genfromtxt.

See what is attached

import os

print("uploads:", os.listdir("/uploads"))

If that list is empty, attach the files and run again.

CSV with a comma delimiter

import os

print("uploads:", os.listdir("/uploads"))
units = np.loadtxt("units.csv", delimiter=",")
prices = np.loadtxt("prices.csv")
print(units)
print("units shape:", units.shape)
print("prices:", prices, prices.shape)

units.csv is 5×3. prices.csv is a single column, so the shape is (3,).

Whitespace .txt

temps = np.loadtxt("temps.txt")
print(temps)
print("shape:", temps.shape)
print("column means (Nairobi, Mombasa, Kisumu):", temps.mean(axis=0))

Default delimiter is any whitespace. temps.txt is 7×3.

Useful loadtxt options

units = np.loadtxt("units.csv", delimiter=",", usecols=(0, 1))
print(units.shape)
print(units)

usecols keeps products A and B.

Missing tokens with genfromtxt

gaps = np.genfromtxt("gaps.csv", delimiter=",", missing_values="nan", filling_values=np.nan)
print(gaps)
print("nan count:", np.isnan(gaps).sum())
print("city nanmean:")
print(np.nanmean(gaps, axis=1))

loadtxt often accepts the token nan too. genfromtxt is the safer default for dirty files.

Pitfall

FileNotFoundError means the workbench cannot see the file in /uploads. Download from the banner, then Add files — the tutorial page cannot put files into the editor for you.