API reference

This document lists the public API surface of bison. Methods marked native are implemented in Mojo. Methods marked stub raise "not implemented" and will be ported in future releases.

DataFrame

Construction

Method	Status
`DataFrame()`	native
`DataFrame.from_pandas(pd_df)`	native
`DataFrame.from_dict(data)`	native
`DataFrame.from_records(records, columns=None)`	native

Properties

Method	Status
`shape()`	native
`size()`	native
`empty()`	native
`columns()`	native
`ndim()`	native
`dtypes()`	native
`info()`	native
`memory_usage(deep=False)`	native

Selection and indexing

Method	Status
`df["col"]`	native
`df[mask]`	native
`df["col"] = series`	native
`get(key, default=null)`	native
`head(n=5)`	native
`tail(n=5)`	native
`iloc`	native
`loc`	native
`at`	native
`iat`	native
`sample(n, frac, replace, random_state)`	native
`select_dtypes(include, exclude)`	native

Aggregation

Method	Status
`sum(axis=0)`	native
`mean(axis=0)`	native
`median(axis=0)`	native (axis=0 only)
`min(axis=0)`	native (axis=0 only)
`max(axis=0)`	native (axis=0 only)
`std(axis=0, ddof=1)`	native (axis=0 only)
`var(axis=0, ddof=1)`	native (axis=0 only)
`count(axis=0)`	native (axis=0 only)
`nunique(axis=0)`	native (axis=0 only)
`quantile(q=0.5, axis=0)`	native (axis=0 only)
`describe(percentiles)`	native

Statistical methods

Method	Status
`sem(axis=0, ddof=1)`	native
`skew(axis=0)`	native
`kurt(axis=0)`	native
`idxmin(axis=0)`	native (axis=0 only)
`idxmax(axis=0)`	native (axis=0 only)
`corr(method="pearson")`	native (pearson only)
`cov(min_periods=1, ddof=1)`	native

Cumulative operations

Method	Status
`cumsum(axis=0)`	native
`cumprod(axis=0)`	native
`cummin(axis=0)`	native
`cummax(axis=0)`	native

Shift and difference

Method	Status
`shift(periods=1)`	native (axis=0 only)
`diff(periods=1)`	native (axis=0 only)
`pct_change(periods=1)`	native (axis=0 only)

Transformations

Method	Status
`abs()`	native
`sqrt()`	native
`exp()`	native
`log()`	native
`log10()`	native
`ceil()`	native
`floor()`	native
`neg()`	native
`round(decimals=0)`	native
`clip(lower, upper)`	native
`astype(dtype)`	native
`where(cond)`	native
`mask(cond)`	native
`isin(values)`	native

Missing data

Method	Status
`isna()` / `isnull()`	native
`notna()` / `notnull()`	native
`fillna(value)`	native
`dropna(axis=0, how="any")`	native
`ffill()`	native
`bfill()`	native
`interpolate(method="linear")`	native

Sorting

Method	Status
`sort_values(by, ascending=True)`	native
`sort_index(ascending=True)`	native

Index operations

Method	Status
`reset_index(drop=False)`	native
`set_index(keys)`	native
`rename(mapper, axis=0)`	native
`rename_axis(mapper, axis=0)`	native
`reindex(new_index)`	native

Dropping and duplicates

Method	Status
`drop(labels, axis=0)`	native
`drop_duplicates(subset, keep="first")`	native
`duplicated(subset, keep="first")`	native

Reshaping

Method	Status
`pivot(index, columns, values)`	native
`pivot_table(values, index, columns, aggfunc)`	native
`melt(id_vars, value_vars)`	native
`stack()`	native
`unstack()`	native
`explode(column)`	native
`transpose()` / `T()`	native
`swaplevel()`	native

Combining

Method	Status
`merge(right, how, on, ...)`	native
`join(other, on, how, ...)`	native
`append(other, ignore_index)`	native
`combine_first(other)`	native
`update(other)`	native

Apply, map, and pipe

Method	Status
`apply(func_name, axis=0)`	native
`apply[F](axis=0)`	native
`applymap(func_name)`	native
`applymap[F]()`	native
`pipe(func_name)`	native
`pipe[F]()`	native
`agg(func)`	stub
`aggregate(func)`	stub
`transform(func)`	stub
`eval(expr)`	native
`query(expr)`	native

Time series

Method	Status
`resample(rule)`	stub
`rolling(window)`	native
`expanding(min_periods)`	native
`ewm(com, span, halflife, alpha)`	native

Output

Method	Status
`to_csv(path)`	native
`to_json(path, orient)`	native
`to_parquet(path)`	native
`to_ipc(path)`	native
`to_excel(path)`	native
`to_dict()`	native
`to_records(index)`	native
`to_numpy()`	native
`to_string()`	native
`to_html()`	native
`to_markdown()`	native
`to_pandas()`	native

GroupBy

Method	Status
`groupby(by)`	native

Series

Construction

Method	Status
`Series()`	native
`Series.from_pandas(pd_s)`	native

Properties

Method	Status
`shape()`	native
`size()`	native
`empty()`	native
`dtype()`	native

Selection

Method	Status
`head(n=5)`	native
`tail(n=5)`	native
`iloc(i)`	native
`at(label)`	native

Arithmetic operators

Method	Status
`add`, `sub`, `mul`, `div`, `floordiv`, `mod`, `pow`	native
`radd`, `rsub`, `rmul`, `rdiv`, `rfloordiv`, `rmod`, `rpow`	native
`__add__`, `__sub__`, `__mul__`, `__truediv__`, etc.	native

Comparison operators

Method	Status
`eq`, `ne`, `lt`, `le`, `gt`, `ge`	native
Scalar comparisons (Float64, Int64, String)	native

Logical operators

Method	Status
`and_`, `or_`, `xor`, `invert`	native
`__and__`, `__or__`, `__xor__`, `__invert__`	native

Aggregation

Method	Status
`sum()`	native
`mean()`	native
`median()`	native
`min()` / `max()`	native
`std(ddof=1)` / `var(ddof=1)`	native
`count()`	native
`nunique()`	native
`quantile(q=0.5)`	native
`describe()`	native
`value_counts(normalize, sort)`	native

Statistical methods

Method	Status
`sem(ddof=1)`	native
`skew()`	native
`kurt()`	native
`idxmin()` / `idxmax()`	native
`corr(other)`	native
`cov(other, ddof=1)`	native

Cumulative operations

Method	Status
`cumsum()`	native
`cumprod()`	native
`cummin()` / `cummax()`	native

Shift and difference

Method	Status
`shift(periods=1)`	native
`diff(periods=1)`	native
`pct_change(periods=1)`	native

Window operations

Method	Status
`rolling(window, min_periods)`	native
`expanding(min_periods)`	native
`ewm(com, span, halflife, alpha)`	native

Missing data

Method	Status
`isna()` / `isnull()`	native
`notna()` / `notnull()`	native
`fillna(value)`	native
`dropna()`	native
`ffill()` / `bfill()`	native

Sorting and ranking

Method	Status
`sort_values(ascending=True)`	native
`sort_index(ascending=True)`	native
`argsort()`	native
`rank()`	native

Transformations

Method	Status
`apply[F]()`	native
`map[F]()`	native
`astype(dtype)`	native
`abs()`	native
`sqrt()`, `exp()`, `log()`, `log10()`	native
`ceil()`, `floor()`, `neg()`	native
`round(decimals=0)`	native
`clip(lower, upper)`	native

Filtering

Method	Status
`unique()`	native
`isin(values)`	native
`between(left, right)`	native
`where(cond)` / `mask(cond)`	native

Output

Method	Status
`to_list()`	native
`to_numpy()`	native
`to_frame(name)`	native
`to_dict()`	native
`to_csv(path)`	native
`to_json(path)`	native
`to_pandas()`	native

Accessors

Method	Status
`str()` -> `StringMethods`	native
`dt()` -> `DatetimeMethods`	native

DataFrameGroupBy

Method	Status
`sum()`	native
`mean()`	native
`median()`	native
`min()` / `max()`	native
`std(ddof=1)` / `var(ddof=1)`	native
`count()`	native
`nunique()`	native
`first()` / `last()`	native
`size()`	native
`agg(func)`	native
`apply(func)`	native
`transform(func)`	native
`filter(func)`	native

SeriesGroupBy

Method	Status
`sum()`	native
`mean()`	native
`min()` / `max()`	native
`std(ddof=1)` / `var(ddof=1)`	native
`count()`	native
`nunique()`	native
`first()` / `last()`	native
`size()`	native
`agg(func)`	native
`apply(func)`	native
`transform(func)`	native

Rolling / SeriesRolling

Returned by DataFrame.rolling(window) / Series.rolling(window). All methods are native.

Method	Description
`sum()`	Rolling sum
`mean()`	Rolling mean
`std(ddof=1)`	Rolling standard deviation
`var(ddof=1)`	Rolling variance
`min()`	Rolling minimum
`max()`	Rolling maximum
`count()`	Rolling count of non-null values

Expanding / SeriesExpanding

Returned by DataFrame.expanding(min_periods) / Series.expanding(min_periods). All methods are native.

Method	Description
`sum()`	Expanding sum
`mean()`	Expanding mean
`std(ddof=1)`	Expanding standard deviation
`var(ddof=1)`	Expanding variance
`min()`	Expanding minimum
`max()`	Expanding maximum
`count()`	Expanding count of non-null values

ExponentialMovingWindow / SeriesExponentialMovingWindow

Returned by DataFrame.ewm() / Series.ewm(). Exactly one of com, span, halflife, or alpha must be provided. All methods are native.

Method	Description
`mean()`	Exponentially weighted moving average
`std(ddof=1)`	Exponentially weighted moving standard deviation
`var(ddof=1)`	Exponentially weighted moving variance

StringMethods (`.str` accessor)

All methods are native.

Method	Description
`upper()`	Convert to uppercase
`lower()`	Convert to lowercase
`strip(to_strip="")`	Strip leading/trailing characters
`lstrip(to_strip="")`	Strip leading characters
`rstrip(to_strip="")`	Strip trailing characters
`contains(pat, regex=False)`	Test if pattern is contained
`startswith(pat)`	Test if starts with pattern
`endswith(pat)`	Test if ends with pattern
`replace(pat, repl, regex=False)`	Replace occurrences
`split(pat, n=-1, expand=False)`	Split strings
`len()`	String length
`find(sub, start=0, end=-1)`	Find substring position
`count(pat)`	Count occurrences
`get(i)`	Get character at position
`slice(start=0, stop, step=1)`	Slice strings
`cat(sep="")`	Concatenate strings
`match(pat)`	Regex match

DatetimeMethods (`.dt` accessor)

All methods are native.

Method	Description
`year()`	Extract year
`month()`	Extract month
`day()`	Extract day
`hour()`	Extract hour
`minute()`	Extract minute
`second()`	Extract second
`dayofweek()`	Day of week (0=Monday)
`dayofyear()`	Day of year
`quarter()`	Quarter (1-4)
`date()`	Date component
`time()`	Time component
`tz_localize(tz)`	Localize timezone
`tz_convert(tz)`	Convert timezone
`floor(freq)`	Floor to frequency
`ceil(freq)`	Ceil to frequency
`round(freq)`	Round to frequency

I/O functions

Function	Status	Notes
`read_csv(filepath, sep, header, ...)`	native	Pure Mojo; dtype inference: bool > int64 > float64 > String
`read_json(path_or_buf, orient, lines)`	native	Supports records, split, columns, index, values, NDJSON
`read_parquet(path, engine, columns)`	native	Via marrow; falls back to pandas for object columns
`read_excel(io, sheet_name, header, ...)`	native	Delegates to pandas (requires openpyxl or xlrd)
`read_ipc(path)`	native	Arrow IPC (Feather v2) via PyArrow
`write_ipc(df, path)`	native	Arrow IPC (Feather v2) via PyArrow
`concat(dfs, axis=0)`	native	Axis=0 with dtype promotion

Data types

BisonDtype constants

Constant	Description
`int8`, `int16`, `int32`, `int64`	Signed integers
`uint8`, `uint16`, `uint32`, `uint64`	Unsigned integers
`float32`, `float64`	Floating point
`bool_`	Boolean
`string_`	String (backed by `List[String]`)
`object_`	Python object
`datetime64_ns`	Datetime (nanosecond precision)
`timedelta64_ns`	Timedelta (nanosecond precision)

Utility functions

Function	Description
`dtype_from_string(name)`	Convert a string like `"int64"` to a `BisonDtype`

Index types

Index

Method	Description
`Index(data, name="")`	Construct from `List[String]`
`len()`	Number of entries
`tolist()`	Convert to `List[String]`
`get_loc(key)`	Find position of a label
`unique()`	Unique values
`sort_values(ascending=True)`	Sort the index

RangeIndex

Method	Description
`RangeIndex(start=0, stop=0, step=1)`	Construct a range index
`len()`	Number of entries

FilesExpand file tree

api-reference.md

Latest commit

History

api-reference.md

File metadata and controls

API reference

DataFrame

Construction

Properties

Selection and indexing

Aggregation

Statistical methods

Cumulative operations

Shift and difference

Transformations

Missing data

Sorting

Index operations

Dropping and duplicates

Reshaping

Combining

Apply, map, and pipe

Time series

Output

GroupBy

Series

Construction

Properties

Selection

Arithmetic operators

Comparison operators

Logical operators

Aggregation

Statistical methods

Cumulative operations

Shift and difference

Window operations

Missing data

Sorting and ranking

Transformations

Filtering

Output

Accessors

DataFrameGroupBy

SeriesGroupBy

Rolling / SeriesRolling

Expanding / SeriesExpanding

ExponentialMovingWindow / SeriesExponentialMovingWindow

StringMethods (.str accessor)

DatetimeMethods (.dt accessor)

I/O functions

Data types

BisonDtype constants

Utility functions

Index types

Index

RangeIndex

StringMethods (`.str` accessor)

DatetimeMethods (`.dt` accessor)