Blog
-
Pandas Series filtering: Use loc with lambda to chain filters
Best filter method for a Python Pandas series? Use loc + lambda, thus filtering the series you got, not the global s:
-
Pandas Series filtering: Chaining loc for multiple conditions
My favorite way to apply multiple filters to a Python Pandas series? Method chaining calling loc 2x: But: This can break! Tomorrow, a better way.
-
Pandas Series filtering: Split long conditions across lines
When applying multiple filters to a Python Pandas series, it’s often best to use multiple lines: This is easier to read, write, and maintain.
-
Pandas boolean filtering: Wrap each condition in parentheses
Want to filter a Python Pandas series with two conditions? One option: Combine two boolean series with &. But careful: Put () around each boolean condition.
-
Pandas boolean indexing: Selecting series items with a mask
Retrieve selected items from a Python Pandas series by broadcasting a condition, then using it as a boolean (“mask”) index:
-
Pandas Series .loc: Selecting with labels and booleans
Retrieve from a Python Pandas series with .loc and indexes: If you pass booleans, True means “return a value,” and False means “ignore it.”
-
New: In-browser Python/Pandas/Git exercises, with AI support
Over my 30+ years teaching Python, I’ve included practice exercises in all of my courses. And when I started to teach Git and Pandas, I made sure to include practice exercises there, too. That’s because there’s no learning without practice. At least, there’s no effective learning. Frustration isn’t fun, but it’s a necessary part of…
-
Pandas Series broadcasting: Arithmetic and comparisons with scalars
Apply an operator to a Python Pandas series, with a scalar value, *broadcasts*: Arithmetic is most obvious: s + 5 # [15, 25, 35]s ** 2 # [100, 400, 900] But comparisons work, too: s >= 20 # [False, True, True]
-
Pandas Series selection speed: iloc[:2] beats loc and head
What’s the fastest way to retrieve the first 2 items from a Python Pandas series?
-
.loc vs .iloc in Pandas: Convenience versus speed
Using .loc to retrieve from a Python Pandas series is more convenient. But .iloc is faster: %timeit s.loc[‘a’] # 2.09 μs%timeit s.iloc[0] # 1.5 μs %timeit s.loc[[‘a’, ‘b’]] # 105 μs%timeit s.iloc[[0, 1]] # 26.4 μs