RacketFrames
This is RacketFrames implemented as a masters thesis at California Polytechnic State University, San Luis Obispo. DataFrames from the popular Pandas library implemented in Racket.
Getting Started
Today procedural and functional programmers alike are working with large amounts of data, and analytical libraries have been developed for many domains to work with such data. Python/Pandas, R or even the command prompt can be used to accomplish these tasks, but this option is especially good if you are 1) using Racket already 2) may need to integrate your solution with other Racket applications in the future or 3) just plain love functional programming.
Prerequisites
Racket Version 6.8 or above.
Installing
Step by step instructions on getting up and running.
Download Racket Version 6.8+ for your OS from the provided prerequisite link.
Documentation
The documentation is a work in progress.
Example Usage
#lang typed/racket
(require RacketFrames)
;******************
;data-frame-mix
;******************
(define columns-mix
(list
(cons 'integer-col (new-ISeries (vector 1 2 3 4)
(build-index-from-labels (list 'a 'b 'c 'd))))
(cons 'categorical-col (new-CSeries (vector 'hello 'world 'fizz 'buzz)))))
; create new data-frame-mix
(define data-frame-mix (new-data-frame columns-mix))
(frame-write-tab data-frame-mix (current-output-port))
; no schema
(define salary-data-frame-csv-no-schema (load-csv-file "sample-csv/salary.csv" #:schema #f #:delim ","))
(data-frame-head salary-data-frame-csv-no-schema)
;; Output ;;
first last age dollar phone
Evan Lloyd 19 $3839.78 (771) 255-1133
Etta Griffith 50 $8158.60 (523) 731-6388
William Conner 50 $9966.70 (759) 504-6619
Rhoda Guerrero 20 $6480.10 (467) 431-4273
Kyle Klein 59 $6106.13 (760) 829-2093
Benjamin Patton 59 $3925.51 (488) 673-5745
Georgie Hansen 51 $8809.92 (579) 706-4402
Gregory Bowen 36 $5176.21 (533) 506-3845
Cornelia Peterson 46 $3626.31 (861) 316-5672
Samuel Cole 37 $7677.20 (760) 406-6331
;; Output ;;
(displayln "DataFrame List of Column Names")
(data-frame-names salary-data-frame-csv-no-schema)
;; Output ;;
DataFrame List of Column Names
'(first last age dollar phone)
;; Output ;;
(displayln "DataFrame Dimensions")
(data-frame-dim salary-data-frame-csv-no-schema)
;; Output ;;
DataFrame Dimensions
(Dim 200 5)
;; Output ;;
(displayln "DataFrame Description")
(show-data-frame-description (data-frame-description salary-data-frame-csv-no-schema))
;; Output ;;
DataFrame Description
DataFrame::(Cols: 5, Rows: 200)
- first: CategoricalSeries
- last: CategoricalSeries
- age: IntegerSeries
- dollar: CategoricalSeries
- phone: CategoricalSeries
;; Output ;;
Running the tests
Unit tests can be run from dataframe/tests which contains individual test files written in Racket using typed/rackunit.
Benchmarks
One of the purposes of this work is to compare performance with other DataFrame implementations. Currently benchmarks are being implemented for comparison against Pandas.
Benchmarks are located in the benchmark directory and separated into different directories by the type of benchmark. They are shells scripts that print the performance of the benchmark against that of Pandas.
Example:
sh benchmark/join_merge/join_merge.sh
Authors
- Shubham Kahal
License
This project is licensed under the MIT License - see the LICENSE.md file for details.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent