Propose adding a delimeter to Column()

[rmasci]

Right now Columns takes an int, and uses strings.Fields to split the line. I'd like to propose we add a delimiter, so that you could parse csv, semicolon, or pipe delimited files. The function would look like this:

func (p *Pipe) Column(col int, d ...string) *Pipe {
	var delim string
	if len(d) > 0 {
		delim = d[0]
	}
	return p.FilterScan(func(line string, w io.Writer) {
		var columns []string
		if delim == "" {
			columns = strings.Fields(line)
		} else {
			columns = strings.Split(line, delim)
		}
		if col > 0 && col <= len(columns) {
			fmt.Fprintln(w, columns[col-1])
		}
	})
}

d is a ...string means it takes 0 or more arguments, so if the user doesn't pass anything it still works maintaining compatibility with existing programs
Ex:

script.Exec(`echo "one two three"`).Columns(2).Stdout()

would still print 2. However

script.Exec(`echo "one,two,three"`).Columns(2, ",").Stdout() 

Would also print 2
Sort of a simple change, and it works for a case where I want to parse through comma separated lines. While I think it works, I might not be thinking of a use case where it would break existing code.

[DioBr4nd0]

Hey @rmasci Can I work on this issue?

[bitfield]

What about something like:

script.File("data.csv").Fields(",").Columns(1, 3, 9).Stdout()

[DioBr4nd0]

sounds nice to me if we can extend it to CSVs along with strings, Should I start working on it ?

[rmasci]

I did this. You can specify the delimiter, the fields like 1,3,7, and the delimiter out. I even added a CSV formatter to convert the CSV to a table like MySQL, grid, or an HTML table. I wanted to ask if it was wanted before issuing a pull request. I keep it in my version of 'script' in my github.

…

On Sat, Sep 7, 2024 at 2:59 PM Rupesh Prajapati ***@***.***> wrote: sounds nice to me if we can extend it to CSVs along with strings — Reply to this email directly, view it on GitHub <#180 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAGRIPKUESWPS5O7PHXBXSDZVNEHRAVCNFSM6AAAAABNXBGB4SVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGMZWGQYDMNBWG4> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[rmasci]

For example: *sample.csv:* Name,Abbreviation,Numeric,Numeric2 Monday,Mon,1,01 Tuesday,Tue,2,02 Wednesday,Wed,3,03 Thursday,Thu,4,04 Friday,Fri,5,05 Saturday,Sat,6,06 Sunday,Sun,7,07 Code: script.File("sample.csv").Fields(",", "|", 1, 3).Stdout() Documentation: func (p *Pipe) Fields(inDelim string, outDelim string, a ...int) *Pipe" *Output:* Name|Numeric Monday|1 Tuesday|2 Wednesday|3 Thursday|4 Friday|5 Saturday|6 Sunday|7 This gives you the option to change the delimiter in the output. I guess 'outDelim' could be eliminated to make the command simpler. 'Feilds(",", 1,3,7,12)' It's sort of like awk -F"," '{print $1,$3,$7,$12}' I also added a 'Table' feature: script.File("sample.csv").Fields(",", ",", 2,4).Table() ``` +-----------------+-------------+ | Abbreviation | Numeric-2 | +============+==========+ | Mon | 01 | | Tue | 02 | | Wed | 03 | | Thu | 04 | | Fri | 05 | | Sat | 06 | | Sun | 07 | +-----------------+--------------+ ``` (looks better in markdown)

…

On Sat, Sep 7, 2024 at 3:25 PM Richard Masci ***@***.***> wrote: I did this. You can specify the delimiter, the fields like 1,3,7, and the delimiter out. I even added a CSV formatter to convert the CSV to a table like MySQL, grid, or an HTML table. I wanted to ask if it was wanted before issuing a pull request. I keep it in my version of 'script' in my github. On Sat, Sep 7, 2024 at 2:59 PM Rupesh Prajapati ***@***.***> wrote: > sounds nice to me if we can extend it to CSVs along with strings > > — > Reply to this email directly, view it on GitHub > <#180 (comment)>, > or unsubscribe > <https://github.com/notifications/unsubscribe-auth/AAGRIPKUESWPS5O7PHXBXSDZVNEHRAVCNFSM6AAAAABNXBGB4SVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGMZWGQYDMNBWG4> > . > You are receiving this because you were mentioned.Message ID: > ***@***.***> >

[bitfield]

Very cool! I'm inspired by nushell to think we should have more data-analysis capabilities like this in script. I don't have any concrete suggestions to make at the moment though. Some use cases would be welcome.

[curio77]

Just my two cents: How about making the delimiter variant a separate Method, like .ColumnsDelim()? And while having a variant taking a rune or string for the second argument is probably nice to have as an easy option, I'd like a .ColumnsRegexp(), too. 🙂

[bitfield]

What about this suggested API, to separate the two concerns of parsing the fields and selecting the fields?

[curio77]

Sounds good, I guess this could/should then have variants for different types of delimiter definition mechanisms (string, regexp, maybe a func)? Speaking of which, having a func option would allow for the regexps I was wishing for to be handled inside that, too, alleviating the need to have a regexp variant.

[rmasci]

I did this in my version of script. I asked at the time if you wanted a pull request but didn't hear anything. This is my version on go.doc:
https://pkg.go.dev/github.com/rmasci/script@v0.9.0#Pipe.Fields

script.File("sample.csv").Fields(",", ",", 2, 4).Stdout() //.Table()

[08:23 PM rmasci@Richs-Mac-mini ~/go/scriptcsv] $ go run main.go
Abbreviation,Numeric-2
Mon,01
Tue,02
Wed,03
Thu,04
Fri,05
Sat,06
Sun,07

If you replacer .Stdout with .Table() function it formats it for you.
https://pkg.go.dev/github.com/rmasci/script@v0.9.0#Pipe.Table

$ go run main.go
+-----------------+--------------+
|   Abbreviation  |   Numeric-2  |
+=================+==============+
|       Mon       |      01      |
|       Tue       |      02      |
|       Wed       |      03      |
|       Thu       |      04      |
|       Fri       |      05      |
|       Sat       |      06      |
|       Sun       |      07      |
+-----------------+--------------+