Skip to content

hygull/node-pandas

BranchesTags

Repository files navigation

node-pandas

An npm package that incorporates minimal features of python pandas. Check it on npm at https://www.npmjs.com/package/node-pandas.

npm NPM

You can also have a look at this colorful documentation at https://hygull.github.io/node-pandas/.

Note: Currently, this package is in development. More methods/functions/attributes will be added with time.

What node-pandas v2.2.0 Can Do

node-pandas brings pandas-like data manipulation to Node.js. Here's what you can do:

Create and manipulate data structures:

  • Create Series from 1D arrays and DataFrames from 2D arrays or CSV files
  • Access data using array-like syntax (indexing, looping, slicing)
  • View data in beautiful tabular format on console
  • Advanced indexing with loc (label-based) and iloc (position-based)

Work with columns and rows:

  • Select specific columns with select()
  • Filter rows with conditions using filter()
  • Access columns by name or index
  • Sort data with sort_values() and sort_index()

Analyze and aggregate data:

  • Group data by columns with groupBy() and aggregate using mean(), sum(), count(), min(), max()
  • Perform statistical analysis on Series and DataFrames
  • Compute cumulative statistics with cumsum(), cumprod(), cummax(), cummin()
  • Calculate rolling and expanding window statistics

Handle missing data:

  • Fill missing values with fillna()
  • Drop missing values with dropna()
  • Detect missing values with isna() and notna()

String operations:

  • Manipulate string data with the str accessor
  • Methods include upper(), lower(), contains(), replace(), split(), and more

Value operations:

  • Get unique values with unique()
  • Count value occurrences with value_counts()
  • Detect and remove duplicates with duplicated() and drop_duplicates()

Comparison operations:

  • Element-wise comparisons with eq(), ne(), gt(), lt(), ge(), le()
  • Range checking with between()

Import and export:

  • Read CSV files with readCsv()
  • Save DataFrames to CSV with toCsv()

Quick Examples:

const pd = require("node-pandas")

// Create a Series
const ages = pd.Series([32, 30, 28])
console.log(ages[0]) // 32

// Create a DataFrame
const df = pd.DataFrame([
    ['Rishikesh Agrawani', 32, 'Engineering'],
    ['Hemkesh Agrawani', 30, 'Marketing'],
    ['Malinikesh Agrawani', 28, 'Sales']
], ['name', 'age', 'department'])

// Select columns
const names = df.select(['name'])

// Filter rows
const over30 = df.filter(row => row.age > 30)

// Group and aggregate
const avgByDept = df.groupBy('department').mean('age')

// Save to CSV
df.toCsv('./output.csv')

Installation

Installation type command
Local npm install node-pandas --save
Local as dev dependency npm install node-pandas --save-dev
Global npm install node-pandas

Table of contents

Series

  1. Example 1 - Creating Series using 1D array/list

  2. Series Methods

DataFrame

  1. Example 1 - Creating DataFrame using 2D array/list

  2. Example 2 - Creating DataFrame using a CSV file

  3. Example 3 - Saving DataFrame in a CSV file

  4. Example 4 - Accessing columns (Retrieving columns using column name) - df.fullName -> ["R A", "B R", "P K"]

  5. Example 5 - Selecting specific columns using select()

  6. Example 6 - Filtering DataFrame rows using filter()

  7. Example 7 - Grouping and aggregating data using groupBy()

  8. Example 8 - Merging DataFrames using merge()

  9. Example 9 - Concatenating DataFrames using concat()

Getting started

Series

Example 1 - Creating Series using 1D array/list

> const pd = require("node-pandas")
undefined
> 
> s = pd.Series([1, 9, 2, 6, 7, -8, 4, -3, 0, 5]) 
NodeSeries [
  1,
  9,
  2,
  6,
  7,
  -8,
  4,
  -3,
  0,
  5,
]
> 
> s.show
┌─────────┬────────┐
│ (index) │ Values │
├─────────┼────────┤
│ 0       │ 1      │
│ 1       │ 9      │
│ 2       │ 2      │
│ 3       │ 6      │
│ 4       │ 7      │
│ 5       │ -8     │
│ 6       │ 4      │
│ 7       │ -3     │
│ 8       │ 0      │
│ 9       │ 5      │
└─────────┴────────┘
undefined
> 
> s[0]  // First element in Series
1
> s.length // Total number of elements 
10
>

Series Methods

Sorting Methods

sort_values()

Sorts Series values in ascending or descending order.

const pd = require("node-pandas")

const s = pd.Series([5, 2, 8, 1, 9], { name: 'numbers' })
console.log(s)
// NodeSeries [ 5, 2, 8, 1, 9 ]

// Sort in ascending order (default)
const sorted_asc = s.sort_values()
console.log(sorted_asc)
// NodeSeries [ 1, 2, 5, 8, 9 ]

// Sort in descending order
const sorted_desc = s.sort_values(false)
console.log(sorted_desc)
// NodeSeries [ 9, 8, 5, 2, 1 ]

sort_index()

Sorts Series by index labels in ascending or descending order.

const pd = require("node-pandas")

const s = pd.Series([10, 20, 30], { index: ['c', 'a', 'b'], name: 'values' })
console.log(s)
// NodeSeries [ 10, 20, 30 ]
// index: ['c', 'a', 'b']

// Sort by index in ascending order
const sorted_asc = s.sort_index()
console.log(sorted_asc)
// NodeSeries [ 20, 30, 10 ]
// index: ['a', 'b', 'c']

// Sort by index in descending order
const sorted_desc = s.sort_index(false)
console.log(sorted_desc)
// NodeSeries [ 10, 30, 20 ]
// index: ['c', 'b', 'a']

Missing Data Handling

fillna()

Fills missing values (null, undefined, NaN) with a specified value.

const pd = require("node-pandas")

const s = pd.Series([1, null, 3, NaN, 5, undefined])
console.log(s)
// NodeSeries [ 1, null, 3, NaN, 5, undefined ]

// Fill missing values with 0
const filled = s.fillna(0)
console.log(filled)
// NodeSeries [ 1, 0, 3, 0, 5, 0 ]

dropna()

Removes all missing values (null, undefined, NaN) from the Series.

const pd = require("node-pandas")

const s = pd.Series([1, null, 3, NaN, 5, undefined])
console.log(s)
// NodeSeries [ 1, null, 3, NaN, 5, undefined ]

// Drop missing values
const cleaned = s.dropna()
console.log(cleaned)
// NodeSeries [ 1, 3, 5 ]

isna()

Returns a boolean Series indicating which values are missing (null, undefined, NaN).

const pd = require("node-pandas")

const s = pd.Series([1, null, 3, NaN, 5])
console.log(s)
// NodeSeries [ 1, null, 3, NaN, 5 ]

// Check for missing values
const missing = s.isna()
console.log(missing)
// NodeSeries [ false, true, false, true, false ]

notna()

Returns a boolean Series indicating which values are not missing.

const pd = require("node-pandas")

const s = pd.Series([1, null, 3, NaN, 5])
console.log(s)
// NodeSeries [ 1, null, 3, NaN, 5 ]

// Check for non-missing values
const notMissing = s.notna()
console.log(notMissing)
// NodeSeries [ true, false, true, false, true ]

Value Operations

unique()

Returns a new Series with unique values, preserving order of first appearance.

const pd = require("node-pandas")

const s = pd.Series([1, 2, 2, 3, 1, 4, 3, 5])
console.log(s)
// NodeSeries [ 1, 2, 2, 3, 1, 4, 3, 5 ]

// Get unique values
const uniqueValues = s.unique()
console.log(uniqueValues)
// NodeSeries [ 1, 2, 3, 4, 5 ]

value_counts()

Returns a Series containing counts of unique values, sorted by frequency in descending order.

const pd = require("node-pandas")

const s = pd.Series(['apple', 'banana', 'apple', 'orange', 'banana', 'apple'])
console.log(s)
// NodeSeries [ 'apple', 'banana', 'apple', 'orange', 'banana', 'apple' ]

// Count occurrences of each value
const counts = s.value_counts()
counts.show
/*
┌─────────┬──────────┬────────┐
│ (index) │ value    │ count  │
├─────────┼──────────┼────────┤
│ 0       │ 'apple'  │ 3      │
│ 1       │ 'banana' │ 2      │
│ 2       │ 'orange' │ 1      │
└─────────┴──────────┴────────┘
*/

duplicated()

Returns a boolean Series indicating duplicate values. The keep parameter controls which duplicates are marked:

  • 'first' (default): Mark duplicates as true except for the first occurrence
  • 'last': Mark duplicates as true except for the last occurrence
  • false: Mark all duplicates as true
const pd = require("node-pandas")

const s = pd.Series([1, 2, 2, 3, 1, 4])
console.log(s)
// NodeSeries [ 1, 2, 2, 3, 1, 4 ]

// Mark duplicates (keep first occurrence)
const isDup = s.duplicated('first')
console.log(isDup)
// NodeSeries [ false, false, true, false, true, false ]

// Mark duplicates (keep last occurrence)
const isDupLast = s.duplicated('last')
console.log(isDupLast)
// NodeSeries [ true, false, true, false, false, false ]

// Mark all duplicates
const isDupAll = s.duplicated(false)
console.log(isDupAll)
// NodeSeries [ true, true, true, false, true, false ]

drop_duplicates()

Returns a new Series with duplicate values removed. The keep parameter controls which duplicates to keep:

  • 'first' (default): Keep the first occurrence
  • 'last': Keep the last occurrence
  • false: Remove all duplicates
const pd = require("node-pandas")

const s = pd.Series([1, 2, 2, 3, 1, 4])
console.log(s)
// NodeSeries [ 1, 2, 2, 3, 1, 4 ]

// Keep first occurrence of duplicates
const uniqueFirst = s.drop_duplicates('first')
console.log(uniqueFirst)
// NodeSeries [ 1, 2, 3, 4 ]

// Keep last occurrence of duplicates
const uniqueLast = s.drop_duplicates('last')
console.log(uniqueLast)
// NodeSeries [ 2, 3, 1, 4 ]

// Remove all duplicates
const noDuplicates = s.drop_duplicates(false)
console.log(noDuplicates)
// NodeSeries [ 3, 4 ]

Comparison Operations

eq()

Element-wise equality comparison. Compares Series values with a scalar or another Series.

const pd = require("node-pandas")

const s = pd.Series([1, 2, 3, 4, 5])
const result = s.eq(3)
console.log(result)
// NodeSeries [ false, false, true, false, false ]

// Compare with another Series
const s1 = pd.Series([1, 2, 3])
const s2 = pd.Series([1, 0, 3])
const result2 = s1.eq(s2)
console.log(result2)
// NodeSeries [ true, false, true ]

ne()

Element-wise not-equal comparison.

const pd = require("node-pandas")

const s = pd.Series([1, 2, 3, 4, 5])
const result = s.ne(3)
console.log(result)
// NodeSeries [ true, true, false, true, true ]

gt()

Element-wise greater-than comparison.

const pd = require("node-pandas")

const s = pd.Series([1, 2, 3, 4, 5])
const result = s.gt(3)
console.log(result)
// NodeSeries [ false, false, false, true, true ]

lt()

Element-wise less-than comparison.

const pd = require("node-pandas")

const s = pd.Series([1, 2, 3, 4, 5])
const result = s.lt(3)
console.log(result)
// NodeSeries [ true, true, false, false, false ]

ge()

Element-wise greater-than-or-equal comparison.

const pd = require("node-pandas")

const s = pd.Series([1, 2, 3, 4, 5])
const result = s.ge(3)
console.log(result)
// NodeSeries [ false, false, true, true, true ]

le()

Element-wise less-than-or-equal comparison.

const pd = require("node-pandas")

const s = pd.Series([1, 2, 3, 4, 5])
const result = s.le(3)
console.log(result)
// NodeSeries [ true, true, true, false, false ]

between()

Check if values fall within a specified range. The inclusive parameter controls boundary inclusion:

  • 'both' (default): Include both boundaries
  • 'neither': Exclude both boundaries
  • 'left': Include left boundary only
  • 'right': Include right boundary only
const pd = require("node-pandas")

const s = pd.Series([1, 2, 3, 4, 5])
const result = s.between(2, 4)
console.log(result)
// NodeSeries [ false, true, true, true, false ]

// Exclude boundaries
const result2 = s.between(2, 4, 'neither')
console.log(result2)
// NodeSeries [ false, false, true, false, false ]

Cumulative Operations

cumsum()

Returns cumulative sum of values. Null values are preserved and skip accumulation.

const pd = require("node-pandas")

const s = pd.Series([1, 2, 3, 4, 5])
const result = s.cumsum()
console.log(result)
// NodeSeries [ 1, 3, 6, 10, 15 ]

// With null values
const s2 = pd.Series([1, null, 3, 4, null, 6])
const result2 = s2.cumsum()
console.log(result2)
// NodeSeries [ 1, null, 4, 8, null, 14 ]

cumprod()

Returns cumulative product of values. Null values are preserved and skip accumulation.

const pd = require("node-pandas")

const s = pd.Series([1, 2, 3, 4, 5])
const result = s.cumprod()
console.log(result)
// NodeSeries [ 1, 2, 6, 24, 120 ]

// With zeros
const s2 = pd.Series([1, 2, 0, 4, 5])
const result2 = s2.cumprod()
console.log(result2)
// NodeSeries [ 1, 2, 0, 0, 0 ]

cummax()

Returns cumulative maximum of values. Null values are preserved and skip accumulation.

const pd = require("node-pandas")

const s = pd.Series([3, 1, 4, 1, 5, 9, 2])
const result = s.cummax()
console.log(result)
// NodeSeries [ 3, 3, 4, 4, 5, 9, 9 ]

// With negative numbers
const s2 = pd.Series([-5, -2, -8, -1, -3])
const result2 = s2.cummax()
console.log(result2)
// NodeSeries [ -5, -2, -2, -1, -1 ]

cummin()

Returns cumulative minimum of values. Null values are preserved and skip accumulation.

const pd = require("node-pandas")

const s = pd.Series([3, 1, 4, 1, 5, 9, 2])
const result = s.cummin()
console.log(result)
// NodeSeries [ 3, 1, 1, 1, 1, 1, 1 ]

// With negative numbers
const s2 = pd.Series([-5, -2, -8, -1, -3])
const result2 = s2.cummin()
console.log(result2)
// NodeSeries [ -5, -5, -8, -8, -8 ]

String Methods

The str accessor provides string manipulation methods that work element-wise on Series values. All methods preserve null values.

str.upper()

Convert strings to uppercase.

const pd = require("node-pandas")

const s = pd.Series(['hello', 'world', null])
const result = s.str.upper()
console.log(result)
// NodeSeries [ 'HELLO', 'WORLD', null ]

str.lower()

Convert strings to lowercase.

const pd = require("node-pandas")

const s = pd.Series(['HELLO', 'WORLD', null])
const result = s.str.lower()
console.log(result)
// NodeSeries [ 'hello', 'world', null ]

str.contains()

Check if strings contain a substring. Optional case-insensitive matching.

const pd = require("node-pandas")

const s = pd.Series(['hello', 'world', null, 'HELLO'])
const result = s.str.contains('ell')
console.log(result)
// NodeSeries [ true, false, null, false ]

// Case-insensitive
const result2 = s.str.contains('ell', false)
console.log(result2)
// NodeSeries [ true, false, null, true ]

str.replace()

Replace occurrences of pattern with replacement string. Supports regex patterns.

const pd = require("node-pandas")

const s = pd.Series(['hello world', 'hello there', null])
const result = s.str.replace('hello', 'hi')
console.log(result)
// NodeSeries [ 'hi world', 'hi there', null ]

str.split()

Split strings by separator and return arrays.

const pd = require("node-pandas")

const s = pd.Series(['a,b,c', 'd,e,f', null])
const result = s.str.split(',')
console.log(result)
// NodeSeries [ ['a','b','c'], ['d','e','f'], null ]

str.strip()

Remove leading and trailing whitespace.

const pd = require("node-pandas")

const s = pd.Series(['  hello  ', '  world', null, 'test  '])
const result = s.str.strip()
console.log(result)
// NodeSeries [ 'hello', 'world', null, 'test' ]

str.startswith()

Check if strings start with a prefix.

const pd = require("node-pandas")

const s = pd.Series(['hello', 'world', null, 'help'])
const result = s.str.startswith('hel')
console.log(result)
// NodeSeries [ true, false, null, true ]

str.endswith()

Check if strings end with a suffix.

const pd = require("node-pandas")

const s = pd.Series(['hello', 'world', null, 'test'])
const result = s.str.endswith('ld')
console.log(result)
// NodeSeries [ false, true, null, false ]

str.len()

Get the length of each string.

const pd = require("node-pandas")

const s = pd.Series(['hello', 'world', null, 'test'])
const result = s.str.len()
console.log(result)
// NodeSeries [ 5, 5, null, 4 ]

Indexing Methods

The loc and iloc accessors provide label-based and position-based indexing for Series data.

loc.get()

Access values by index labels. Supports single labels and arrays of labels.

const pd = require("node-pandas")

const s = pd.Series([10, 20, 30, 40], { index: ['a', 'b', 'c', 'd'] })
console.log(s)
// NodeSeries [ 10, 20, 30, 40 ]
// index: ['a', 'b', 'c', 'd']

// Get single value by label
const value = s.loc.get('b')
console.log(value)
// 20

// Get multiple values by labels
const values = s.loc.get(['a', 'c', 'd'])
console.log(values)
// NodeSeries [ 10, 30, 40 ]
// index: ['a', 'c', 'd']

iloc.get()

Access values by integer positions. Supports single positions and arrays of positions.

const pd = require("node-pandas")

const s = pd.Series([10, 20, 30, 40], { index: ['a', 'b', 'c', 'd'] })
console.log(s)
// NodeSeries [ 10, 20, 30, 40 ]

// Get single value by position
const value = s.iloc.get(1)
console.log(value)
// 20

// Get multiple values by positions
const values = s.iloc.get([0, 2, 3])
console.log(values)
// NodeSeries [ 10, 30, 40 ]

loc.set()

Set values by index labels. Supports single labels and arrays of labels.

const pd = require("node-pandas")

const s = pd.Series([10, 20, 30, 40], { index: ['a', 'b', 'c', 'd'] })

// Set single value by label
s.loc.set('b', 99)
console.log(s)
// NodeSeries [ 10, 99, 30, 40 ]

// Set multiple values by labels
s.loc.set(['a', 'c'], [100, 300])
console.log(s)
// NodeSeries [ 100, 99, 300, 40 ]

iloc.set()

Set values by integer positions. Supports single positions and arrays of positions.

const pd = require("node-pandas")

const s = pd.Series([10, 20, 30, 40], { index: ['a', 'b', 'c', 'd'] })

// Set single value by position
s.iloc.set(1, 99)
console.log(s)
// NodeSeries [ 10, 99, 30, 40 ]

// Set multiple values by positions
s.iloc.set([0, 2], [100, 300])
console.log(s)
// NodeSeries [ 100, 99, 300, 40 ]

Window Operations

Window operations allow you to perform calculations over sliding or expanding windows of data.

rolling()

Create a rolling window for calculating statistics over a fixed window size.

const pd = require("node-pandas")

const s = pd.Series([1, 2, 3, 4, 5, 6, 7, 8, 9, 10])

// Rolling mean with window size 3
const rollingMean = s.rolling(3).mean()
console.log(rollingMean)
// NodeSeries [ null, null, 2, 3, 4, 5, 6, 7, 8, 9 ]

// Rolling sum with window size 3
const rollingSum = s.rolling(3).sum()
console.log(rollingSum)
// NodeSeries [ null, null, 6, 9, 12, 15, 18, 21, 24, 27 ]

// Rolling min with window size 3
const rollingMin = s.rolling(3).min()
console.log(rollingMin)
// NodeSeries [ null, null, 1, 2, 3, 4, 5, 6, 7, 8 ]

// Rolling max with window size 3
const rollingMax = s.rolling(3).max()
console.log(rollingMax)
// NodeSeries [ null, null, 3, 4, 5, 6, 7, 8, 9, 10 ]

// Rolling standard deviation with window size 3
const rollingStd = s.rolling(3).std()
console.log(rollingStd)
// NodeSeries [ null, null, 1, 1, 1, 1, 1, 1, 1, 1 ]

expanding()

Create an expanding window that includes all values from the start up to the current position.

const pd = require("node-pandas")

const s = pd.Series([1, 2, 3, 4, 5])

// Expanding mean
const expandingMean = s.expanding().mean()
console.log(expandingMean)
// NodeSeries [ 1, 1.5, 2, 2.5, 3 ]

// Expanding sum
const expandingSum = s.expanding().sum()
console.log(expandingSum)
// NodeSeries [ 1, 3, 6, 10, 15 ]

// Expanding min
const expandingMin = s.expanding().min()
console.log(expandingMin)
// NodeSeries [ 1, 1, 1, 1, 1 ]

// Expanding max
const expandingMax = s.expanding().max()
console.log(expandingMax)
// NodeSeries [ 1, 2, 3, 4, 5 ]

// Expanding standard deviation
const expandingStd = s.expanding().std()
console.log(expandingStd)
// NodeSeries [ 0, 0.707..., 1, 1.29..., 1.58... ]

DataFrame

Example 1 - Creating DataFrame using 2D array/list

> const pd = require("node-pandas")
undefined
> 
> columns = ['full_name', 'user_id', 'technology']
[ 'full_name', 'user_id', 'technology' ]
> 
> df = pd.DataFrame([
...     ['Guido Van Rossum', 6, 'Python'],
...     ['Ryan Dahl', 5, 'Node.js'],
...     ['Anders Hezlsberg', 7, 'TypeScript'],
...     ['Wes McKinney', 3, 'Pandas'],
...     ['Ken Thompson', 1, 'B language']
... ], columns)
NodeDataFrame [
  [ 'Guido Van Rossum', 6, 'Python' ],
  [ 'Ryan Dahl', 5, 'Node.js' ],
  [ 'Anders Hezlsberg', 7, 'TypeScript' ],
  [ 'Wes McKinney', 3, 'Pandas' ],
  [ 'Ken Thompson', 1, 'B language' ],
  columns: [ 'full_name', 'user_id', 'technology' ],
  index: [ 0, 1, 2, 3, 4 ],
  rows: 5,
  cols: 3,
  out: true
]
> 
> df.show
┌─────────┬────────────────────┬─────────┬──────────────┐
│ (index) │ full_name          │ user_id │ technology   │
├─────────┼────────────────────┼─────────┼──────────────┤
│ 0       │ 'Guido Van Rossum' │ 6       │ 'Python'     │
│ 1       │ 'Ryan Dahl'        │ 5       │ 'Node.js'    │
│ 2       │ 'Anders Hezlsberg' │ 7       │ 'TypeScript' │
│ 3       │ 'Wes McKinney'     │ 3       │ 'Pandas'     │
│ 4       │ 'Ken Thompson'     │ 1       │ 'B language' │
└─────────┴────────────────────┴─────────┴──────────────┘
undefined
> 
> df.index
[ 0, 1, 2, 3, 4 ]
> 
> df.columns
[ 'full_name', 'user_id', 'technology' ]
>

Example 2 - Creating DataFrame using a CSV file

Note: If CSV will have multiple newlines b/w 2 consecutive rows, no problem, it takes care of it and considers as single newline.

df = pd.readCsv(csvPath) where CsvPath is absolute/relative path of the CSV file.

Examples:

df = pd.readCsv("../node-pandas/docs/csvs/devs.csv")

df = pd.readCsv("/Users/hygull/Projects/NodeJS/node-pandas/docs/csvs/devs.csv")

devs.csv » cat /Users/hygull/Projects/NodeJS/node-pandas/docs/csvs/devs.csv

fullName,Profession,Language,DevId
Ken Thompson,C developer,C,1122
Ron Wilson,Ruby developer,Ruby,4433
Jeff Thomas,Java developer,Java,8899


Rishikesh Agrawani,Python developer,Python,6677
Kylie Dwine,C++,C++ Developer,0011

Briella Brown,JavaScript developer,JavaScript,8844

Now have a look the below statements executed on Node REPL.

> const pd = require("node-pandas")
undefined
> 
> df = pd.readCsv("/Users/hygull/Projects/NodeJS/node-pandas/docs/csvs/devs.csv")
NodeDataFrame [
  {
    fullName: 'Ken Thompson',
    Profession: 'C developer',
    Language: 'C',
    DevId: 1122
  },
  {
    fullName: 'Ron Wilson',
    Profession: 'Ruby developer',
    Language: 'Ruby',
    DevId: 4433
  },
  {
    fullName: 'Jeff Thomas',
    Profession: 'Java developer',
    Language: 'Java',
    DevId: 8899
  },
  {
    fullName: 'Rishikesh Agrawani',
    Profession: 'Python developer',
    Language: 'Python',
    DevId: 6677
  },
  {
    fullName: 'Kylie Dwine',
    Profession: 'C++',
    Language: 'C++ Developer',
    DevId: 11
  },
  {
    fullName: 'Briella Brown',
    Profession: 'JavaScirpt developer',
    Language: 'JavaScript',
    DevId: 8844
  },
  columns: [ 'fullName', 'Profession', 'Language', 'DevId' ],
  index: [ 0, 1, 2, 3, 4, 5 ],
  rows: 6,
  cols: 4,
  out: true
]
> 
> df.index
[ 0, 1, 2, 3, 4, 5 ]
> 
> df.columns
[ 'fullName', 'Profession', 'Language', 'DevId' ]
> 
> df.show
┌─────────┬──────────────────────┬────────────────────────┬─────────────────┬───────┐
│ (index) │ fullName             │ Profession             │ Language        │ DevId │
├─────────┼──────────────────────┼────────────────────────┼─────────────────┼───────┤
│ 0       │ 'Ken Thompson'       │ 'C developer'          │ 'C'             │ 1122  │
│ 1       │ 'Ron Wilson'         │ 'Ruby developer'       │ 'Ruby'          │ 4433  │
│ 2       │ 'Jeff Thomas'        │ 'Java developer'       │ 'Java'          │ 8899  │
│ 3       │ 'Rishikesh Agrawani' │ 'Python developer'     │ 'Python'        │ 6677  │
│ 4       │ 'Kylie Dwine'        │ 'C++'                  │ 'C++ Developer' │ 11    │
│ 5       │ 'Briella Brown'      │ 'JavaScript developer' │ 'JavaScript'    │ 8844  │
└─────────┴──────────────────────┴────────────────────────┴─────────────────┴───────┘
undefined
> 
> df[0]['fullName']
'Ken Thompson'
> 
> df[3]['Profession']
'Python developer'
> 
> df[5]['Language']
'JavaScript'
>

Example 3 - Saving DataFrame in a CSV file

Note: Here we will save DataFrame in /Users/hygull/Desktop/newDevs.csv (in this case) which can be different in your case.

> const pd = require("node-pandas")
undefined
> 
> df = pd.readCsv("./docs/csvs/devs.csv")
NodeDataFrame [
  {
    fullName: 'Ken Thompson',
    Profession: 'C developer',
    Language: 'C',
    DevId: 1122
  },
  {
    fullName: 'Ron Wilson',
    Profession: 'Ruby developer',
    Language: 'Ruby',
    DevId: 4433
  },
  {
    fullName: 'Jeff Thomas',
    Profession: 'Java developer',
    Language: 'Java',
    DevId: 8899
  },
  {
    fullName: 'Rishikesh Agrawani',
    Profession: 'Python developer',
    Language: 'Python',
    DevId: 6677
  },
  {
    fullName: 'Kylie Dwine',
    Profession: 'C++',
    Language: 'C++ Developer',
    DevId: 11
  },
  {
    fullName: 'Briella Brown',
    Profession: 'JavaScirpt developer',
    Language: 'JavaScript',
    DevId: 8844
  },
  columns: [ 'fullName', 'Profession', 'Language', 'DevId' ],
  index: [ 0, 1, 2, 3, 4, 5 ],
  rows: 6,
  cols: 4,
  out: true
]
> 
> df.cols
4
> df.rows
6
> df.columns
[ 'fullName', 'Profession', 'Language', 'DevId' ]
> df.index
[ 0, 1, 2, 3, 4, 5 ]
> 
> df.toCsv("/Users/hygull/Desktop/newDevs.csv")
undefined
> CSV file is successfully created at /Users/hygull/Desktop/newDevs.csv

>

Let's see content of /Users/hygull/Desktop/newDevs.csv

cat /Users/hygull/Desktop/newDevs.csv

fullName,Profession,Language,DevId
Ken Thompson,C developer,C,1122
Ron Wilson,Ruby developer,Ruby,4433
Jeff Thomas,Java developer,Java,8899
Rishikesh Agrawani,Python developer,Python,6677
Kylie Dwine,C++,C++ Developer,11
Briella Brown,JavaScript developer,JavaScript,8844

Example 4 - Accessing columns (Retrieving columns using column name)

CSV file (devs.csv): ./docs/csvs/devs.csv

const pd = require("node-pandas")
df = pd.readCsv("./docs/csvs/devs.csv") // Node DataFrame object

df.show // View DataFrame in tabular form
/*
┌─────────┬──────────────────────┬────────────────────────┬─────────────────┬───────┐
│ (index) │ fullName             │ Profession             │ Language        │ DevId │
├─────────┼──────────────────────┼────────────────────────┼─────────────────┼───────┤
│ 0       │ 'Ken Thompson'       │ 'C developer'          │ 'C'             │ 1122  │
│ 1       │ 'Ron Wilson'         │ 'Ruby developer'       │ 'Ruby'          │ 4433  │
│ 2       │ 'Jeff Thomas'        │ 'Java developer'       │ 'Java'          │ 8899  │
│ 3       │ 'Rishikesh Agrawani' │ 'Python developer'     │ 'Python'        │ 6677  │
│ 4       │ 'Kylie Dwine'        │ 'C++'                  │ 'C++ Developer' │ 11    │
│ 5       │ 'Briella Brown'      │ 'JavaScirpt developer' │ 'JavaScript'    │ 8844  │
└─────────┴──────────────────────┴────────────────────────┴─────────────────┴───────┘
*/

console.log(df['fullName'])
/*
    NodeSeries [
      'Ken Thompson',
      'Ron Wilson',
      'Jeff Thomas',
      'Rishikesh Agrawani',
      'Kylie Dwine',
      'Briella Brown'
    ]
*/

console.log(df.DevId)
/* 
    NodeSeries [ 1122, 4433, 8899, 6677, 11, 8844 ]
*/

let languages = df.Language
console.log(languages) 
/*
    NodeSeries [
      'C',
      'Ruby',
      'Java',
      'Python',
      'C++ Developer',
      'JavaScript'
    ]
*/

console.log(languages[0], '&', languages[1]) // C & Ruby


let professions = df.Profession
console.log(professions) 
/*
    NodeSeries [
      'C developer',
      'Ruby developer',
      'Java developer',
      'Python developer',
      'C++',
      'JavaScirpt developer'
    ]
*/

// Iterate like arrays
for(let profession of professions) {
    console.log(profession)
}
/*
    C developer
    Ruby developer
    Java developer
    Python developer
    C++
    JavaScirpt developer
*/

Example 5 - Selecting specific columns using select()

Note: The select() method returns a new DataFrame containing only the specified columns.

const pd = require("node-pandas")

// Create a DataFrame with employee data
const df = pd.DataFrame([
    ['Rishikesh Agrawani', 32, 'Engineering'],
    ['Hemkesh Agrawani', 30, 'Marketing'],
    ['Malinikesh Agrawani', 28, 'Sales']
], ['name', 'age', 'department'])

df.show
/*
┌─────────┬──────────────────────┬─────┬──────────────┐
│ (index) │ name                 │ age │ department   │
├─────────┼──────────────────────┼─────┼──────────────┤
│ 0       │ 'Rishikesh Agrawani' │ 32  │ 'Engineering'│
│ 1       │ 'Hemkesh Agrawani'   │ 30  │ 'Marketing'  │
│ 2       │ 'Malinikesh Agrawani'│ 28  │ 'Sales'      │
└─────────┴──────────────────────┴─────┴──────────────┘
*/

// Select a single column
const nameOnly = df.select(['name'])
nameOnly.show
/*
┌─────────┬──────────────────────┐
│ (index) │ name                 │
├─────────┼──────────────────────┤
│ 0       │ 'Rishikesh Agrawani' │
│ 1       │ 'Hemkesh Agrawani'   │
│ 2       │ 'Malinikesh Agrawani'│
└─────────┴──────────────────────┘
*/

// Select multiple columns
const nameAndAge = df.select(['name', 'age'])
nameAndAge.show
/*
┌─────────┬──────────────────────┬─────┐
│ (index) │ name                 │ age │
├─────────┼──────────────────────┼─────┤
│ 0       │ 'Rishikesh Agrawani' │ 32  │
│ 1       │ 'Hemkesh Agrawani'   │ 30  │
│ 2       │ 'Malinikesh Agrawani'│ 28  │
└─────────┴──────────────────────┴─────┘
*/

// Original DataFrame remains unchanged
console.log(df.columns) // ['name', 'age', 'department']

Example 6 - Filtering DataFrame rows using filter()

Note: The filter() method returns a new DataFrame containing only rows that match the condition. Multiple filters can be chained together.

const pd = require("node-pandas")

// Create a DataFrame with employee data
const df = pd.DataFrame([
    ['Rishikesh Agrawani', 32, 'Engineering'],
    ['Hemkesh Agrawani', 30, 'Marketing'],
    ['Malinikesh Agrawani', 28, 'Sales']
], ['name', 'age', 'department'])

df.show
/*
┌─────────┬──────────────────────┬─────┬──────────────┐
│ (index) │ name                 │ age │ department   │
├─────────┼──────────────────────┼─────┼──────────────┤
│ 0       │ 'Rishikesh Agrawani' │ 32  │ 'Engineering'│
│ 1       │ 'Hemkesh Agrawani'   │ 30  │ 'Marketing'  │
│ 2       │ 'Malinikesh Agrawani'│ 28  │ 'Sales'      │
└─────────┴──────────────────────┴─────┴──────────────┘
*/

// Filter rows where age is greater than 28
const over28 = df.filter(row => row.age > 28)
over28.show
/*
┌─────────┬──────────────────────┬─────┬──────────────┐
│ (index) │ name                 │ age │ department   │
├─────────┼──────────────────────┼─────┼──────────────┤
│ 0       │ 'Rishikesh Agrawani' │ 32  │ 'Engineering'│
│ 1       │ 'Hemkesh Agrawani'   │ 30  │ 'Marketing'  │
└─────────┴──────────────────────┴─────┴──────────────┘
*/

// Filter rows where department is 'Engineering'
const engineering = df.filter(row => row.department === 'Engineering')
engineering.show
/*
┌─────────┬──────────────────────┬─────┬──────────────┐
│ (index) │ name                 │ age │ department   │
├─────────┼──────────────────────┼─────┼──────────────┤
│ 0       │ 'Rishikesh Agrawani' │ 32  │ 'Engineering'│
└─────────┴──────────────────────┴─────┴──────────────┘
*/

// Chain multiple filters together
const result = df
    .filter(row => row.age > 28)
    .filter(row => row.department !== 'Sales')
result.show
/*
┌─────────┬──────────────────────┬─────┬──────────────┐
│ (index) │ name                 │ age │ department   │
├─────────┼──────────────────────┼─────┼──────────────┤
│ 0       │ 'Rishikesh Agrawani' │ 32  │ 'Engineering'│
│ 1       │ 'Hemkesh Agrawani'   │ 30  │ 'Marketing'  │
└─────────┴──────────────────────┴─────┴──────────────┘
*/

Example 7 - Grouping and aggregating data using groupBy()

Note: The groupBy() method groups rows by one or more columns and allows aggregation using methods like mean(), sum(), count(), min(), and max().

const pd = require("node-pandas")

// Create a DataFrame with employee data including departments
const df = pd.DataFrame([
    ['Rishikesh Agrawani', 32, 'Engineering', 95000],
    ['Hemkesh Agrawani', 30, 'Marketing', 75000],
    ['Malinikesh Agrawani', 28, 'Sales', 65000],
    ['Alice Johnson', 29, 'Engineering', 92000],
    ['Bob Smith', 31, 'Marketing', 78000],
    ['Carol White', 27, 'Sales', 62000]
], ['name', 'age', 'department', 'salary'])

df.show
/*
┌─────────┬──────────────────────┬─────┬──────────────┬────────┐
│ (index) │ name                 │ age │ department   │ salary │
├─────────┼──────────────────────┼─────┼──────────────┼────────┤
│ 0       │ 'Rishikesh Agrawani' │ 32  │ 'Engineering'│ 95000  │
│ 1       │ 'Hemkesh Agrawani'   │ 30  │ 'Marketing'  │ 75000  │
│ 2       │ 'Malinikesh Agrawani'│ 28  │ 'Sales'      │ 65000  │
│ 3       │ 'Alice Johnson'      │ 29  │ 'Engineering'│ 92000  │
│ 4       │ 'Bob Smith'          │ 31  │ 'Marketing'  │ 78000  │
│ 5       │ 'Carol White'        │ 27  │ 'Sales'      │ 62000  │
└─────────┴──────────────────────┴─────┴──────────────┴────────┘
*/

// Single-column grouping: Group by department and calculate mean salary
const avgSalaryByDept = df.groupBy('department').mean('salary')
avgSalaryByDept.show
/*
┌─────────┬──────────────┬──────────────┐
│ (index) │ department   │ salary_mean  │
├─────────┼──────────────┼──────────────┤
│ 0       │ 'Engineering'│ 93500        │
│ 1       │ 'Marketing'  │ 76500        │
│ 2       │ 'Sales'      │ 63500        │
└─────────┴──────────────┴──────────────┘
*/

// Group by department and calculate sum of salaries
const totalSalaryByDept = df.groupBy('department').sum('salary')
totalSalaryByDept.show
/*
┌─────────┬──────────────┬──────────────┐
│ (index) │ department   │ salary_sum   │
├─────────┼──────────────┼──────────────┤
│ 0       │ 'Engineering'│ 187000       │
│ 1       │ 'Marketing'  │ 153000       │
│ 2       │ 'Sales'      │ 127000       │
└─────────┴──────────────┴──────────────┘
*/

// Group by department and count employees
const countByDept = df.groupBy('department').count()
countByDept.show
/*
┌─────────┬──────────────┬───────┐
│ (index) │ department   │ count │
├─────────┼──────────────┼───────┤
│ 0       │ 'Engineering'│ 2     │
│ 1       │ 'Marketing'  │ 2     │
│ 2       │ 'Sales'      │ 2     │
└─────────┴──────────────┴───────┘
*/

// Group by department and find minimum age
const minAgeByDept = df.groupBy('department').min('age')
minAgeByDept.show
/*
┌─────────┬──────────────┬──────────┐
│ (index) │ department   │ age_min  │
├─────────┼──────────────┼──────────┤
│ 0       │ 'Engineering'│ 29       │
│ 1       │ 'Marketing'  │ 30       │
│ 2       │ 'Sales'      │ 27       │
└─────────┴──────────────┴──────────┘
*/

// Group by department and find maximum age
const maxAgeByDept = df.groupBy('department').max('age')
maxAgeByDept.show
/*
┌─────────┬──────────────┬──────────┐
│ (index) │ department   │ age_max  │
├─────────┼──────────────┼──────────┤
│ 0       │ 'Engineering'│ 32       │
│ 1       │ 'Marketing'  │ 31       │
│ 2       │ 'Sales'      │ 28       │
└─────────┴──────────────┴──────────┘
*/

// Multi-column grouping: Group by department and age range
const groupedByDeptAndAge = df.groupBy(['department', 'age']).count()
groupedByDeptAndAge.show
/*
┌─────────┬──────────────┬─────┬───────┐
│ (index) │ department   │ age │ count │
├─────────┼──────────────┼─────┼───────┤
│ 0       │ 'Engineering'│ 29  │ 1     │
│ 1       │ 'Engineering'│ 32  │ 1     │
│ 2       │ 'Marketing'  │ 30  │ 1     │
│ 3       │ 'Marketing'  │ 31  │ 1     │
│ 4       │ 'Sales'      │ 27  │ 1     │
│ 5       │ 'Sales'      │ 28  │ 1     │
└─────────┴──────────────┴─────┴───────┘
*/

Example 8 - Merging DataFrames using merge()

Note: The merge() method combines two DataFrames based on a join key, supporting inner, left, right, and outer joins.

const pd = require("node-pandas")

// Create two DataFrames to merge
const df1 = pd.DataFrame([
    [1, 'Rishikesh Agrawani'],
    [2, 'Hemkesh Agrawani'],
    [3, 'Malinikesh Agrawani']
], ['id', 'name'])

const df2 = pd.DataFrame([
    [1, 25],
    [2, 30],
    [3, 35]
], ['id', 'age'])

// Inner join on id column
const merged = df1.merge(df2, 'id', 'inner')
merged.show
/*
┌─────────┬────┬──────────────────────┬─────┐
│ (index) │ id │ name                 │ age │
├─────────┼────┼──────────────────────┼─────┤
│ 0       │ 1  │ 'Rishikesh Agrawani' │ 25  │
│ 1       │ 2  │ 'Hemkesh Agrawani'   │ 30  │
│ 2       │ 3  │ 'Malinikesh Agrawani'│ 35  │
└─────────┴────┴──────────────────────┴─────┘
*/

// Left join - keeps all rows from left DataFrame
const leftMerged = df1.merge(df2, 'id', 'left')
leftMerged.show
/*
┌─────────┬────┬──────────────────────┬─────┐
│ (index) │ id │ name                 │ age │
├─────────┼────┼──────────────────────┼─────┤
│ 0       │ 1  │ 'Rishikesh Agrawani' │ 25  │
│ 1       │ 2  │ 'Hemkesh Agrawani'   │ 30  │
│ 2       │ 3  │ 'Malinikesh Agrawani'│ 35  │
└─────────┴────┴──────────────────────┴─────┘
*/

Example 9 - Concatenating DataFrames using concat()

Note: The concat() method stacks DataFrames vertically (axis=0) or horizontally (axis=1).

const pd = require("node-pandas")

// Create DataFrames to concatenate
const df1 = pd.DataFrame([
    [1, 'Rishikesh Agrawani'],
    [2, 'Hemkesh Agrawani']
], ['id', 'name'])

const df2 = pd.DataFrame([
    [3, 'Malinikesh Agrawani']
], ['id', 'name'])

// Vertical concatenation (stack rows)
const verticalConcat = pd.DataFrame.concat([df1, df2], 0)
verticalConcat.show
/*
┌─────────┬────┬──────────────────────┐
│ (index) │ id │ name                 │
├─────────┼────┼──────────────────────┤
│ 0       │ 1  │ 'Rishikesh Agrawani' │
│ 1       │ 2  │ 'Hemkesh Agrawani'   │
│ 2       │ 3  │ 'Malinikesh Agrawani'│
└─────────┴────┴──────────────────────┘
*/

// Horizontal concatenation (stack columns)
const df3 = pd.DataFrame([
    [25, 'Engineering'],
    [30, 'Marketing']
], ['age', 'department'])

const horizontalConcat = pd.DataFrame.concat([df1, df3], 1)
horizontalConcat.show
/*
┌─────────┬────┬──────────────────────┬─────┬──────────────┐
│ (index) │ id │ name                 │ age │ department   │
├─────────┼────┼──────────────────────┼─────┼──────────────┤
│ 0       │ 1  │ 'Rishikesh Agrawani' │ 25  │ 'Engineering'│
│ 1       │ 2  │ 'Hemkesh Agrawani'   │ 30  │ 'Marketing'  │
└─────────┴────┴──────────────────────┴─────┴──────────────┘
*/

References

About

An npm package that incorporates minimal features of python pandas.

www.npmjs.com/package/node-pandas

Topics

nodejs javascript pandas node-pandas

Resources

Readme

License

MIT license
Activity

Stars

23 stars

Watchers

1 watching

Forks

9 forks
Report repository

Releases

5 tags

Packages

 
 
 

Contributors

Languages