csvutil

csvutil

Command line utility for transforming CSV files.

Design Choices

• Human-readable. Operations and arguments should be easy to remember. There is no shorthand syntax.
• Safe. External programs that can properly escape shell strings should be able to spawn csvutil processes safely.
• Not a database. All operations are done in memory and there are no fancy tricks to speed things up.

Install

Csvutil may be installed directly from the source repository on a system with Go 1.16 or higher, and may be done like so:

$ go install github.com/evantbyrne/csvutil@latest

Note: The $ at the beginning of newlines in this document represents the bash shell prompt, and is not a part of the actual commands.

Usage

$ csvutil source1.csv --operation1 --operation2 source2.csv --operation1 ...

There are two types of arguments that may be passed to csvutil:

1. Sources. These are CSV files to read from.
2. Operations denoted with the -- prefix. These transform the current source.

Upon success, csvutil prints the resulting CSV data to STDOUT.

Examples

Here are a few examples of working with csvutil.

Reading files

$ csvutil user.csv
username,id,group_id
foo,1,10
bar,2,
baz,3,20
four,4,20
five,5,
$ csvutil group.csv
name,id
Admin,10
Moderator,20

Filtering

$ csvutil user.csv --where username == baz --or group_id != 20
username,id,group_id
foo,1,10
bar,2,
baz,3,20
five,5,

Join

$ csvutil user.csv group.csv --join group_id == id
username,id,group_id,name,id
foo,1,10,Admin,10
baz,3,20,Moderator,20
four,4,20,Moderator,20

Sort and select

$ csvutil test/user.csv --sort username ALPHA DESC --select id,username
id,username
4,four
1,foo
5,five
3,baz
2,bar

Operations

Concat

Appends one source onto another.

No arguments.

$ csvutil foo.csv
id,name
1,Foo
2,Bar
$ csvutil bar.csv
id,name
3,Baz
$ csvutil foo.csv bar.csv --concat
id,name
1,Foo
2,Bar
3,Baz

Count

Returns the total number of rows for the current source.

Arguments:

• Column name.

$ csvutil user.csv --count total
total
5

Distinct

Removes duplicates from the current source.

Arguments:

• Comma-separated list of columns that must match to be considered duplicate. Use "*" to compare all columns.

$ csvutil foo.csv
id,name,group_id
1,Foo,10
2,Bar,10
3,Foo,20
4,Foo,10
5,Foobar,20
1,Baz,20
$ csvutil foo.csv --distinct name
id,name,group_id
1,Foo,10
2,Bar,10
5,Foobar,20
1,Baz,20
$ csvutil foo.csv --distinct name,group_id
id,name,group_id
1,Foo,10
2,Bar,10
3,Foo,20
5,Foobar,20
1,Baz,20

Except

Remove rows from the previous source if they appear in the current source.

Arguments:

• Comma-separated list of columns that must match to be considered duplicate. Use "*" to compare all columns.

$ csvutil foo.csv
id,name
1,Foo
2,Bar
3,Baz
4,Foobar
$ csvutil bar.csv
id,name
2,Two
3,Three
5,Five
6,Six
$ csvutil foo.csv bar.csv --except id
id,name
1,Foo
4,Foobar

Join

Perform an inner join on two sources.

Arguments:

• Column from previous source to compare.
• Operation. Choices: ==, !=
• Column from the current source to compare.

$ csvutil user.csv
username,id,group_id
foo,1,10
bar,2,
baz,3,20
four,4,20
five,5,
$ csvutil group.csv
name,id
Admin,10
Moderator,20
$ csvutil user.csv group.csv --join group_id == id
username,id,group_id,name,id
foo,1,10,Admin,10
baz,3,20,Moderator,20
four,4,20,Moderator,20

Select

Choose a subset of columns from the current source.

Arguments:

• Comma-separated list of columns to select.

$ csvutil foo.csv
id,name,group_id
1,Foo,10
2,Bar,10
3,Baz,20
$ csvutil foo.csv --select name,id
name,id
Foo,1
Bar,2
Baz,3

Sort

Reorder the current source.

Arguments:

• Column to compare.
• Algorithm. Choices: ALPHA, FLOAT, INT
• Order. Choices: ASC, DESC

$ csvutil user.csv
username,id,group_id
foo,1,10
bar,2,
baz,3,20
four,4,20
five,5,
$ csvutil user.csv --sort username ALPHA ASC
username,id,group_id
bar,2,
baz,3,20
five,5,
foo,1,10
four,4,20

When using FLOAT and INT, values that cannot be converted will be considered lower than values which may be converted, and they will be compared to each other alphabetically. e.g.,

$ csvutil baz.csv
year,cost
10,500
2,200.10
2000,50.01
x,x
2019,1000.99
y,z
z,y
$ csvutil baz.csv --sort year INT DESC
year,cost
2019,1000.99
2000,50.01
10,500
2,200.10
z,y
y,z
x,x

Values

Strip the header from the current source.

$ csvutil user.csv --count total
total
5
$ csvutil user.csv --count total --values
5

Where

Filter the current source.

Arguments:

• Column name to compare.
• Operator. Choices: ==, !=, IN, NOT_IN
• Value to look for.

$ csvutil user.csv
username,id,group_id
foo,1,10
bar,2,
baz,3,20
four,4,20
five,5,
$ csvutil user.csv --where username == baz --or group_id != 20
username,id,group_id
foo,1,10
bar,2,
baz,3,20
five,5,

The IN and NOT_IN operators take column-separated lists of values. e.g.,

$ csvutil user.csv --where id IN 1,3
username,id,group_id
foo,1,10
baz,3,20

Operations may be chained to perform "and" conditionals. e.g.,

$ csvutil user.csv --where username != baz --where id != 2
username,id,group_id
foo,1,10
four,4,20
five,5,