Insyra With Me

How to do your first data analysis in Go with Insyra

By Tim Lai · · 3 min read · 閱讀中文版

Guides & Use Cases#getting-started#datatable#csv#ccl#groupby

Goal

By the end of this post you’ll have one small Go program that does the classic first analysis: load a CSV, add a revenue column, total it per region, sort it, and save the answer as a new CSV. No notebook, no second language. Just go run.

Prerequisites

  • Go 1.26 or newer. Insyra v0.3.5 sets go 1.26.9 in its go.mod. With Go 1.21+ the go command downloads that toolchain for you unless GOTOOLCHAIN=local is set.
  • A fresh module:
mkdir first-insyra && cd first-insyra
go mod init first-insyra
go get github.com/HazelnutParadise/insyra/allpkgs

Everything below was run with Insyra v0.3.5.

Step 1: Load the CSV into a DataTable

Our data is six orders. To keep the example copy-and-run, the CSV sits in a string; with a real file you’d call insyra.ReadCSV_FileWithOptions("sales.csv", opts) and pass the same options.

const salesCSV = `region,product,price,qty
North,Coffee,120,3
South,Tea,80,5
North,Tea,80,2
East,Coffee,120,4
South,Coffee,120,1
East,Tea,80,6
`

dt, err := insyra.ReadCSV_StringWithOptions(salesCSV, insyra.CSVReadOptions{
	FirstRowToColNames: true,
})
if err != nil {
	panic(err)
}
dt.ShowHead(3)

FirstRowToColNames turns the header row into column names. ShowHead(3) prints the first three rows, so you can see what you loaded (and what type each column became) before doing anything else.

Step 2: Add a column with CCL

CCL, the Column Calculation Language, works like Excel formulas. You can refer to a column by its letter (A, B, …) or by its name in ['...']:

dt.AddColUsingCCL("revenue", "['price'] * ['qty']")

That’s the whole step. Every row now has revenue = price × qty.

Step 3: Group, sum and sort

byRegion := dt.GroupBy("region").Aggregate(
	insyra.AggregateConfig{SourceCol: "revenue", Op: insyra.OpSum, As: "total_revenue"},
)
byRegion.SortBy(insyra.DataTableSortConfig{ColumnName: "total_revenue", Descending: true})
byRegion.Show()

GroupBy takes one or more key columns, and Aggregate takes one config per output column. OpSum is one of several operations; OpMean, OpMedian, OpCount and friends work the same way.

Step 4: Pull one column out as a DataList

A column of a DataTable is a DataList, and a DataList knows its own statistics:

revenue := dt.GetColByName("revenue")
fmt.Println("Mean revenue per order:", revenue.Mean())

Step 5: Save the result

if err := byRegion.ToCSV("revenue_by_region.csv", false, true, false); err != nil {
	panic(err)
}

The three booleans are: write row names as the first column (no), write column names as the first row (yes), add a UTF-8 BOM (no; turn it on if the file is headed for Excel).

The whole program

package main

import (
	"fmt"

	"github.com/HazelnutParadise/insyra"
)

const salesCSV = `region,product,price,qty
North,Coffee,120,3
South,Tea,80,5
North,Tea,80,2
East,Coffee,120,4
South,Coffee,120,1
East,Tea,80,6
`

func main() {
	dt, err := insyra.ReadCSV_StringWithOptions(salesCSV, insyra.CSVReadOptions{
		FirstRowToColNames: true,
	})
	if err != nil {
		panic(err)
	}
	dt.ShowHead(3)

	dt.AddColUsingCCL("revenue", "['price'] * ['qty']")

	byRegion := dt.GroupBy("region").Aggregate(
		insyra.AggregateConfig{SourceCol: "revenue", Op: insyra.OpSum, As: "total_revenue"},
	)
	byRegion.SortBy(insyra.DataTableSortConfig{ColumnName: "total_revenue", Descending: true})
	byRegion.Show()

	revenue := dt.GetColByName("revenue")
	fmt.Println("Mean revenue per order:", revenue.Mean())

	if err := byRegion.ToCSV("revenue_by_region.csv", false, true, false); err != nil {
		panic(err)
	}
}

Expected result

go run . prints this (Insyra v0.3.5):

DataTable (6 rows x 4 columns) [showing rows 0 to 2]
================================================================================
 stat (selected range): C(mean=93.33, range=[80.00, 120.00]) 
--------------------------------------------------------------------------------
RowNames   A(region)  B(product)  C(price)  D(qty) 
---------- ---------- ----------- --------- -------
0:         'North'    'Coffee'         120       3 
1:         'South'    'Tea'             80       5 
2:         'North'    'Tea'             80       2 

DataTable (3 rows x 2 columns)
================================================================================
 stat: B(mean=666.67, range=[520.00, 960.00]) 
--------------------------------------------------------------------------------
RowNames   A(region)  B(total_revenue) 
---------- ---------- -----------------
0:         'East'                  960 
1:         'North'                 520 
2:         'South'                 520 

Mean revenue per order: 333.3333333333333

And revenue_by_region.csv contains:

region,total_revenue
East,960
North,520
South,520

Small thing I like here: Show puts a quick stat line on top of the table (mean and range of the numeric column), so a weird number jumps out before you’ve written a single check.

Next steps