Incomplete sheet

This sheet is incomplete and could use some attention. Please submit code snippet suggestions as an issue or PR here.

Pandas Series

Opinionated words of caution:

• Types are (currently?) a mess, and autocasting does not make this any better.
• Specify dtype when defining series to avoid surprises further down the pipeline.
• Missing values are a complete mess: None/Null/NA/NaN are all used interchangeably, despite the existence of is.na and is.null implying otherwise.
• The representation of missing values differs per type (float uses nan even for NA inputs, object uses NA), which makes the whole thing even more confusing.
• This seems to be fixed with the new Int64 and Float64 types, although this won't help in practice as autocasting still uses the old types.
• Avoid indexes unless you have a good reason. Consider using a DataFrame instead.
• The in operator, counterintuitively, works on the index, not the values.

Code

import pandas as pd

Create

Action	Code	Details
From DataFrame column s	data[s]

Create empty series

Action	Code	Details
Empty series (of object type)	pd.Series()
Empty boolean series	pd.Series(dtype=bool)
Empty categorical series with pre-defined categories	pd.Categorical([], categories=['a', 'b', 'c'])
Empty categorical series without defined categories	pd.Series(dtype='category')
Empty int series	pd.Series(dtype=int)
Empty float series	pd.Series(dtype=float)
Empty datetime series	pd.Series(dtype='datetime64[ns]')

Create series of constant values

Action	Code	Details
Series filled with NAs of length n (of object type)	pd.Series([None] * n)
Constant value v of length n	pd.Series(v, index=range(n))

Create series from a list of values

Action	Code	Details
Object series from a generic list of values	pd.Series([1, None, 'a'])
Int series from a list of integers	pd.Series([1, '2', 3], dtype='int')
Nullable int series from a list of integers	pd.Series([1, None, 3], dtype = 'Int64')	Preserves None as
Nullable int series filled with NA of length n	pd.Series([None] * 3, dtype='Int64')
Float series from list of numbers	pd.Series([1, None, 3], dtype='float')	None is converted to NaN!
Nullable float series from a list of numbers	pd.Series([1, None, 3.5], dtype = 'Float64')	None, NA and NaN are all set to
Nullable float series filled with NA of length n	pd.Series([None] * 3, dtype='Int64')
Categorical series from list of strings	pd.Categorical(['b', 'b', 'a'], categories=['a', 'b', 'c'])
Categorical series filled with NA of length n	pd.Categorical([None] * n, categories=['a', 'b', 'c'])

Test

Action	Code	Details
Is series or subclass	isinstance(x, pd.Series)
Is series and not subclass	type(x) is pd.Series
Empty	x.empty
Not empty	not x.empty
Has length n	len(x) == n
Is boolean series	pd.api.types.is_bool_dtype(x)
Is categorical series	isinstance(x.dtype, pd.CategoricalDtype)
Is ordered categorical series	isinstance(x.dtype, pd.CategoricalDtype) and x.ordered
Is numeric series	pd.api.types.is_numeric_dtype(x)	For example, int or float
Is integer series	pd.api.types.is_integer_dtype(x)	For example, int or Int64
Is unsigned integer series	pd.api.types.is_unsigned_integer_dtype(x)
Is float series	pd.api.types.is_float_dtype(x)	For example, float or Float64
Is datetime64 series	pd.api.types.is_datetime64_dtype(x)
Is datetime64[ns] series	pd.api.types.is_datetime64_ns_dtype(x)
Is string series	pd.api.types.is_string_dtype(x)
Is object series	pd.api.types.is_object_dtype(x)
Is hashable series	pd.api.types.is_hashable(x)
No duplicate elements (all unique)	x.is_unique
Any duplicate elements	not x.is_unique
Contains NA	x.hasnans
Contains only NA	x.isna().all()
Contains no NA	x.notna().all()
Contains value v, ignoring NAs	x.isin([v]).any()	WARNING: `v in x' tests the indices instead!
Contains value v, ignoring NAs	any(x == v)
Contains any of the values v1, v2, ignoring NAs	x.isin([v1, v2]).any()
Does not contain value v	all(x != v)
Are elements in increasing order	x.is_monotonic_increasing
Are elements in decreasing order	x.is_monotonic_decreasing

Test boolean series

Action	Code	Details
All values are True	x.all()
All values are False	not x.any()

Tests for string series

Action	Code	Details
Contains string s	x.str.contains(s).any()

Assertions

Action	Code	Details
Assert series equal	pd.testing.assert_series_equal(x, y)
Assert series equal, ignoring the names	pd.testing.assert_series_equal(x, y, check_names=False)

Extract

Action	Code	Details
Number of elements	x.size
Number of elements	len(x)
Hash	hash(x.values.tobytes())
Dtype	x.dtype
Number of unique elements, ignoring NAs	x.nunique()
Smallest value, ignoring NAs	x.min()
Index of the smallest value, ignoring NAs	x.idxmin()
Greatest value, ignoring NAs	x.max()
Index of the greatest value, ignoring NAs	x.idxmax()
Count occurrence per value	x.value_counts()

Dtype-specific operations

Action	Code	Details
Get length of each list element	x.list.len()
Get the _i_th item of each list element	x.list[i]

Update

Warning: updates may change the dtype of the series!

Action	Code	Details
Set element at index i to NA	x[i] = pd.NA
Set element at index i to value v	x[i] = v

Derive

Cast

Action	Code	Details
Cast series to numeric dtype	pd.to_numeric(x)

Order

Action	Code	Details
Reverse	x[::-1]
Sort ascendingly	x.sort_values()	Nulls are placed last
Sort descendingly	x.sort_values(ascending=False)	Nulls are placed last
Shuffle	x.sample(frac=1)

Transform

Action	Code	Details
Fill NA by value_v_	x.fillna(value=v)
Replace value v by NA	x.replace(V, pd.NA)	May cast series to another type if nulls are not supported (example: int)

Mask

Action	Code	Details
Duplicate mask	x.duplicated()

Cast to type

Action	Code	Details
Cast to boolean	x.astype(bool)
Cast to integer	x.astype(int)
Cast to float	x.astype(float)
Cast to string	x.astype(str)
Cast to categorical (nominal)	x.astype('category')
Cast to categorical (nominal) with given categories	pd.Categorical(x, categories=['a', 'b', 'c'])
Cast to ordered categorical (ordinal)	pd.Categorical(x, ordered=True)
Cast categorical to integer codes	pd.Series(z.codes)
Cast ordered categorical to unordered categorical	x.as_unordered()

Grow

Shrink

Action	Code	Details
First n elements	x.head(n)
Last n elements	x.tail(n)
Slice	x[a:b]
Sample n elements	x.sample(n)
Remove duplicates	x.drop_duplicates()

Convert

Action	Code	Details
To list	x.tolist()
To list	list(x)
To set of unique values	set(x)	NA is included only once, nan is only included once
To DataFrame (single column)	x.to_frame()
To Numpy ND array	x.to_numpy()
To dict (index-value pairs)	x.to_dict()
To JSON index-value array	x.to_json()