To obtain 10 random records from your DataFrame `df`, you would use:

df.sample(10) df.random(10) df.head(10) df.tail(10) `head()` or `tail()` will give you the first 10 or last 10 records, respectively, and are therefore not random. `sample()` will provide 10 records at random. There is no function called `random()`.

Does the `corr()` function, when applied to a DataFrame, derive correlation coefficient values for all numerical columns while ignoring non-numerical columns?

True False Yes, this is true. However, be careful about the numerical values in a column. If a column has zip codes or lat/lon location values, `corr()` will also derive correlations with the numerical values of those columns, which makes no sense in many situations. It may not be a bad idea to hand-pick the columns for which you want to derive correlations.

Exercise