/joinery

Data frames for Java

Primary LanguageJavaGNU General Public License v3.0GPL-3.0

joinery

joinery [joi-nuh-ree]
1. In woodworking, the craft of joining together pieces of wood to produce more complex items.
2. In Java, a data analysis library for joining together pieces of data to produce insight.

Build Status Codecov Maven Central

quick start

Remember FizzBuzz (of course you do!), well imagine you have just solved the puzzle (well done!) and you have written the results to a comma-delimited file for further analysis. Now you want to know how many times are the strings Fizz, Buzz, and FizzBuzz printed out.

You could answer this question any number of ways, for example you could modify the original program, or reach for Python/pandas, or even (for the sadistic among us, you know who you are) type out a one-liner at the command prompt (probably including cut, sort, and uniq).

Well, now you have one more option. This option is especially good if you are 1) using Java already and 2) may need to integrate your solution with other Java applications in the future.

You can answer this question with joinery.

df.groupBy("value")
  .count()
  .sortBy("-number")
  .head(3)

Printing out the resulting data frame gives us the following table.

  	   value 	number
 0	Fizz    	    27
 1	Buzz    	    14
 2	FizzBuzz	     6

See FizzBuzz.java for the complete code.

next steps

Get the executable jar and try it for yourself.

$ java -jar joinery-dataframe-1.10-jar-with-dependencies.jar shell
# Joinery -- Data frames for Java, 1.10-deb702e
# OpenJDK 64-Bit Server VM, Oracle Corporation, 1.8.0_92-internal
# Rhino 1.7 release 2 2009 03 22
> df = new DataFrame()
[empty data frame]
> df.add("value")
[empty data frame]
> [10, 20, 30].forEach(function(val) {
      df.append([val])
  })
> df
        value
 0	   10
 1	   20
 2	   30

>

maven

Since version 1.10, joinery is included in the central repo. If you are upgrading from a prior version, note the new group id.

<dependency>
  <groupId>sh.joinery</groupId>
  <artifactId>joinery-dataframe</artifactId>
  <version>1.10</version>
</dependency>

utilities

joinery includes some tools to make working with data frames easier. These tools are available by running joinery.DataFrame as an application.

$ java joinery.DataFrame
usage: joinery.DataFrame [compare|plot|show|shell] [csv-file ...]

show

Show displays the tabular data of a data frame in a gui window.

$ java joinery.DataFrame show data.csv

Screenshot of show window

plot

Plot displays the numeric data of a data frame as a chart.

$ java joinery.DataFrame plot data.csv

Screenshot of plot window

shell

Launches an interactive JavaScript shell for working with data frames.

$ java joinery.DataFrame shell
# Joinery -- Data frames for Java, 1.10-deb702e
# OpenJDK 64-Bit Server VM, Oracle Corporation, 1.8.0_92-internal
# Rhino 1.7 release 2 2009 03 22
> df = DataFrame.readCsv("https://www.quandl.com/api/v1/datasets/GOOG/NASDAQ_AAPL.csv")
              Date	  Open	  High	   Low	        Close	             Volume
    0	2015-03-20	128.25	128.4	125.16	 125.90000000	  68695136.00000000
    1	2015-03-19	128.75	129.25	127.4	 127.50000000	  45809490.00000000
    2	2015-03-18	127.0	129.16	126.37	 128.47000000	  65270945.00000000
    3	2015-03-17	125.9	127.32	125.65	 127.04000000	  51023104.00000000
    4	2015-03-16	123.88	124.95	122.87	 124.95000000	  35874300.00000000
    5	2015-03-13	124.4	125.4	122.58	 123.59000000	  51827283.00000000
    6	2015-03-12	122.31	124.9	121.63	 124.45000000	  48362719.00000000
    7	2015-03-11	124.75	124.77	122.11	 122.24000000	  68938974.00000000
    8	2015-03-10	126.41	127.22	123.8	 124.51000000	  68856582.00000000

... 8649 rows skipped ...

 8658	1980-12-12	0.0	4.12	4.11	   4.11000000	  14657300.00000000

> df.types()
[class java.util.Date, class java.lang.String, class java.lang.String, class java.lang.String, class java.lang.Double, class java.lang.Double]
> df.sortBy("Date")
              Date     Open     High     Low            Close                Volume
 8658	1980-12-12	0.0	4.12	4.11	   4.11000000	  14657300.00000000
 8657	1980-12-15	0.0	3.91	3.89	   3.89000000	   5496400.00000000
 8656	1980-12-16	0.0	3.62	3.61	   3.61000000	   3304000.00000000
 8655	1980-12-17	0.0	3.71	3.7 	   3.70000000	   2701300.00000000
 8654	1980-12-18	0.0	3.82	3.8 	   3.80000000	   2295300.00000000
 8653	1980-12-19	0.0	4.05	4.04	   4.04000000	   1519700.00000000
 8652	1980-12-22	0.0	4.25	4.23	   4.23000000	   1167600.00000000
 8651	1980-12-23	0.0	4.43	4.41	   4.41000000	   1467200.00000000
 8650	1980-12-24	0.0	4.66	4.64	   4.64000000	   1500100.00000000

... 8649 rows skipped ...

    0	2015-03-20	128.25	128.4	125.16	 125.90000000	  68695136.00000000

> .reindex("Date")
	       Open	High	 Low	        Close	             Volume
1980-12-12	0.0	4.12	4.11	   4.11000000	  14657300.00000000
1980-12-15	0.0	3.91	3.89	   3.89000000	   5496400.00000000
1980-12-16	0.0	3.62	3.61	   3.61000000	   3304000.00000000
1980-12-17	0.0	3.71	3.7 	   3.70000000	   2701300.00000000
1980-12-18	0.0	3.82	3.8 	   3.80000000	   2295300.00000000
1980-12-19	0.0	4.05	4.04	   4.04000000	   1519700.00000000
1980-12-22	0.0	4.25	4.23	   4.23000000	   1167600.00000000
1980-12-23	0.0	4.43	4.41	   4.41000000	   1467200.00000000
1980-12-24	0.0	4.66	4.64	   4.64000000	   1500100.00000000

... 8649 rows skipped ...

2015-03-20	128.25	128.4	125.16	 125.90000000	  68695136.00000000

> .retain("Close")
	                Close
1980-12-12	   4.11000000
1980-12-15	   3.89000000
1980-12-16	   3.61000000
1980-12-17	   3.70000000
1980-12-18	   3.80000000
1980-12-19	   4.04000000
1980-12-22	   4.23000000
1980-12-23	   4.41000000
1980-12-24	   4.64000000

... 8649 rows skipped ...

2015-03-20	 125.90000000

> .plot(PlotType.AREA)

documentation

The complete api documentation for the DataFrame class is available at https://joinery.sh