csvmung is a simple tool to manipulate csv files. It can perform the following transformations:

• Filter out rows that don't match a regular expression.
• Output only rows with unique values.
• Output only certain columns.
• Split columns based on a regular expression.

You can use the following transformations. All column references are zero indexed.

re:<column>:<pattern>

Only pass through rows where the column matches the pattern.

split:<column>:<pattern>

Split column into separate columns using pattern.

uniq:<column>

Only pass through rows where the value in column hasn't yet been seen.

cols:<col1>:<col2>:...:<colN>

Only pass through the specified columns. If the column specifier isn't an integer, the value will be passed through in that position.

Given a csv file with the following contents, named foo.csv:

a,b-j,c,d
a,k-n,l,m
a,n-f,g,x
a,p-p,d,d

You can get all the rows that contain 'd' in the fourth column:

$ csvmung -i foo.csv re:3:'^d$'
a,b-j,c,d
a,p-p,d,d

or only output the columns with a unique value in the fourth column:

$ csvmung -i foo.csv uniq:3
a,b-j,c,d
a,k-n,l,m
a,n-f,g,x

or output just the first and fourth columns:

$ csvmung -i foo.csv cols:0:3
a,d
a,m
a,x
a,d

or split the value in the second column:

$ csvmung -i foo.csv split:1:'-'
a,b,j,c,d
a,k,n,l,m
a,n,f,g,x
a,p,p,d,d

Finally, you can chain the operators so that each works on the output of the last:

$ csvmung -i foo.csv split:1:'-' re:4:'^d$'
a,b,j,c,d
a,p,p,d,d