DataFrame

new DataFrame(data, options?)

Construct a DataFrame from an object of arrays, an array of objects, another DataFrame, or a 2D array.

Parameters

dataRecord<string, unknown[]> | object[] | DataFrame | unknown[][]

The data.

options?{ index?, columns? }

Construction options.

Returns

DataFrame

Example

import { DataFrame } from 'pandas-ts';
const df = new DataFrame({ a: [1, 2, 3], b: [4, 5, 6] });
df.shape;    // => [3, 2]

df.shape

Return a [rows, columns] tuple.

Returns

[number, number]

Example

new DataFrame({ x: [1,2,3], y: [4,5,6] }).shape;    // => [3, 2]

df.columns

Return the column labels of the DataFrame.

Returns

string[]

Example

new DataFrame({ a: [1], b: [2], c: [3] }).columns;    // => ['a', 'b', 'c']

df.index

Return the row labels of the DataFrame.

Returns

unknown[]

Example

new DataFrame({ a: [1,2,3] }).index;    // => [0, 1, 2]

df.dtypes

Return a Series with the dtype of each column.

Returns

Record<string, DType>

Example

const df = new DataFrame({ a: [1,2], b: ['x','y'] });
df.dtypes;    // => { a: DType('int64'), b: DType('object') }

df.head(n?)

Return the first n rows. Default n is 5.

Returns

DataFrame

Example

const df = new DataFrame({ a: [1,2,3,4,5,6] });
df.head(3).shape;    // => [3, 1]

df.tail(n?)

Return the last n rows.

Returns

DataFrame

Example

new DataFrame({ a: [1,2,3,4,5,6] }).tail(2).shape;    // => [2, 1]

df.describe()

Generate descriptive statistics for numeric columns: count, mean, std, min, 25%, 50%, 75%, max.

Returns

DataFrame

Example

const df = new DataFrame({ a: [1,2,3,4,5] });
df.describe().index;    // => ['count','mean','std','min','25%','50%','75%','max']

df.col(name)

Select a single column by name, returning it as a Series.

Parameters

namestring

Column name.

Returns

Series

Example

const df = new DataFrame({ a: [1,2,3], b: [4,5,6] });
df.col('a').to_list();    // => [1, 2, 3]

df.assign(cols)

Assign new columns to a DataFrame. Returns a new DataFrame with all existing columns plus any new ones.

Parameters

colsRecord<string, Series | unknown[]>

Columns to assign.

Returns

DataFrame

Example

const df = new DataFrame({ a: [1,2,3] });
df.assign({ b: df.col('a').mul(2) }).columns;    // => ['a', 'b']

df.drop(labels, options?)

Drop specified labels from rows or columns.

Parameters

labelsstring | string[]

Labels to drop.

options?{ axis? }

axis=0 (rows) or axis=1 (columns). Default 0.

Returns

DataFrame

Example

const df = new DataFrame({ a: [1,2], b: [3,4], c: [5,6] });
df.drop(['b'], { axis: 1 }).columns;    // => ['a', 'c']

df.rename(mapping)

Rename columns using a mapping.

Returns

DataFrame

Example

const df = new DataFrame({ a: [1,2], b: [3,4] });
df.rename({ a: 'alpha', b: 'beta' }).columns;    // => ['alpha', 'beta']

df.fillna(value)

Fill NA/NaN values.

Returns

DataFrame

Example

const df = new DataFrame({ a: [1, NaN, 3] });
df.fillna(0).col('a').to_list();    // => [1, 0, 3]

df.dropna(options?)

Remove missing values.

Returns

DataFrame

Example

const df = new DataFrame({ a: [1, NaN, 3], b: [4, 5, NaN] });
df.dropna().shape;    // => [1, 2]

df.groupby(by, options?)

Group the DataFrame by one or more columns.

Returns

DataFrameGroupBy

Example

const df = new DataFrame({ city: ['A','B','A'], sales: [100, 200, 150] });
df.groupby('city').sum().col('sales').to_list();    // => [250, 200]

df.merge(right, options?)

Merge DataFrame objects with a database-style join.

Parameters

rightDataFrame

Right DataFrame to merge.

options?{ on?, how?, left_on?, right_on? }

Merge options. how defaults to 'inner'.

Returns

DataFrame

Example

const left = new DataFrame({ id: [1,2], name: ['Alice','Bob'] });
const right = new DataFrame({ id: [1,2], score: [90, 85] });
left.merge(right, { on: 'id' }).col('score').to_list();    // => [90, 85]

df.pivot_table(options)

Create a spreadsheet-style pivot table as a DataFrame.

Returns

DataFrame

Example

const df = new DataFrame({
  city: ['A','A','B','B'],
  year: [2023,2024,2023,2024],
  sales: [100,120,200,180]
});
const pt = df.pivot_table({ index: 'city', columns: 'year', values: 'sales', aggfunc: 'sum' });
pt.shape;    // => [2, 2]

df.melt(options?)

Unpivot a DataFrame from wide to long format.

Returns

DataFrame

Example

const df = new DataFrame({ id: [1,2], jan: [100,200], feb: [110,210] });
const melted = df.melt({ id_vars: 'id', value_vars: ['jan','feb'], var_name: 'month', value_name: 'sales' });
melted.shape;    // => [4, 3]

df.sort_values(by, options?)

Sort by the values along either axis.

Parameters

bystring | string[]

Column(s) to sort by.

options?{ ascending? }

Sort direction.

Returns

DataFrame

Example

const df = new DataFrame({ a: [3,1,2], b: ['c','a','b'] });
df.sort_values('a').col('a').to_list();    // => [1, 2, 3]

df.query(expr)

Query the columns of a DataFrame with a boolean expression string.

Returns

DataFrame

Example

const df = new DataFrame({ a: [1,2,3,4,5], b: [10,20,30,40,50] });
df.query('a > 3').col('a').to_list();    // => [4, 5]

df.evalExpr(expr)

Evaluate an expression against the DataFrame columns. Returns a Series with the result. (Note: named evalExpr because eval is reserved in JavaScript.)

Returns

Series

Example

const df = new DataFrame({ a: [1,2,3], b: [4,5,6] });
df.evalExpr('a + b').to_list();    // => [5, 7, 9]

df.corr(method?)

Compute pairwise Pearson correlation of columns.

Returns

DataFrame

Example

const df = new DataFrame({ a: [1,2,3,4], b: [2,4,6,8] });
df.corr().col('a').iloc(1);    // => 1  (perfect correlation)

df.cov_matrix(min_periods?, ddof?)

Compute pairwise covariance of columns (powered by numera).

Returns

DataFrame

Example

const df = new DataFrame({ a: [1,2,3,4,5] });
df.cov_matrix().col('a').iloc(0);    // => 2.5

df.to_csv(path?, options?)

Write DataFrame to a CSV file or return as a string.

Returns

string | void

Example

const df = new DataFrame({ a: [1,2], b: [3,4] });
df.to_csv();    // => 'a,b\n1,3\n2,4\n'

df.to_json(path?, options?)

Convert the DataFrame to a JSON string.

Returns

string | void

Example

new DataFrame({ a: [1,2] }).to_json();
// => '[{"a":1},{"a":2}]'

df.T

Transpose the DataFrame (swap rows and columns).

Returns

DataFrame

Example

const df = new DataFrame({ a: [1,2], b: [3,4] });
df.T.shape;    // => [2, 2]

df.sum(axis?)

Sum of values. When axis=0 (default) returns a column-wise record. When axis=1 returns a row-wise Series.

Returns

Record<string,number> | Series

Example

const df = new DataFrame({ a: [1,2], b: [3,4] });
df.sum(1).to_list();    // => [4, 6]

df.mean(axis?)

Mean of values. When axis=1 returns a row-wise Series.

Returns

Record<string,number> | Series

Example

df.mean(1).to_list();    // => [2, 3]

df.dropna({ axis?, how?, thresh? })

Remove missing values. axis=1 drops columns (instead of rows) that contain NA values according to how/thresh.

Returns

DataFrame

Example

const df = new DataFrame({ a: [1, null], b: [3, 4] });
df.dropna({ axis: 1 }).columns;    // => ['b']

DataFrame.concat(dfs, { axis? })

Concatenate DataFrames. axis=1 stacks them side-by-side (column-wise). Duplicate column names are auto-suffixed with _1, _2, etc.

Returns

DataFrame

Example

const df1 = new DataFrame({ a: [1,2] });
const df2 = new DataFrame({ b: [3,4] });
DataFrame.concat([df1, df2], { axis: 1 }).columns;    // => ['a','b']

df.to_dict(orient?)

Convert the DataFrame to a dictionary. orient controls the output shape: 'dict' (default), 'list', 'records', 'series', 'split', 'index'.

Returns

object

Example

const df = new DataFrame({ a: [1,2], b: [3,4] });
df.to_dict('records');    // => [{ a:1, b:3 }, { a:2, b:4 }]
df.to_dict('split');      // => { index:[0,1], columns:['a','b'], data:[[1,3],[2,4]] }

df.to_string({ index?, max_rows?, max_cols? })

Render the DataFrame to a tab-separated string suitable for console output. Truncates to max_rows / max_cols with ... indicators.

Returns

string

Example

const df = new DataFrame({ x: [1,2,3], y: [4,5,6] });
console.log(df.to_string({ max_rows: 10 }));

df.convert_dtypes()

Convert columns to the best possible dtype. String/object columns whose values are all numeric strings are promoted to float64.

Returns

DataFrame

Example

const df = new DataFrame({ a: ['1','2','3'], b: ['x','y','z'] });
df.convert_dtypes().col('a').dtype.name;    // => 'float64'

df.corr_matrix(method?, min_periods?)

Compute pairwise correlation. Supports 'pearson' (default) and 'spearman' (rank-based). Returns a square DataFrame.

Returns

DataFrame

Example

const df = new DataFrame({ a: [1,2,3,4], b: [4,3,2,1] });
df.corr_matrix('spearman').col('b').iloc(0);    // => -1