Select starting sheet in html/xls/sqlite from command-line

[aborruso]

Hi,
if I use vd -b t.html -o html.csv I have the table below in CSV and not the html table I have inside my file.

tag,id,nrows,ncols,classes
table,,72,4,wikitable sortable

table is the html table inside my t.html file. Is there a way to pass to the command line the sheet name? Something like vd -b t.html -sheet table -o html.csv

Thank you

[saulpw]

Hi @aborruso, there's not an easy way yet. I've wanted something like this myself at times. Let me see if I can come up with something. Thanks for the suggestion!

[aborruso]

Hi @saulpw is there a way to open the table directly, when it is only one?

My final goal is to use VisiData as HTML to CSV converter with something like below in which I use an xpath query to extract only one table

curl "http://example.com/page.html" | myScrapeUtilty -xpathRule '//table[count(tr/td)>7]' | vd -b  -f html -o out.csv

But also with only one table visidata asks me to choose, and it saves as csv sheets sheet

Thank you

[saulpw]

Hi @aborruso, try adding -p dive.vd with the attached small .vd script.

sheet	col	row	longname	input	keystrokes	comment
			open-file	-	o	
-		0	dive-row		^J	

The first command opens the input from stdin (-), and the second command dives into the first row (0).

You can get this .vd yourself with:

1. the same command you have but without -b
2. press Enter and do any other manual steps
3. press Shift+D to go to the commandlog
4. finally, press Ctrl+S to save to dive.vd, which you can use with your pipeline.

dive.vd.txt

[aborruso]

@saulpw you are really brilliant, I'm impressed VisiData is a kind of magic

[aborruso]

@saulpw I have added a recipe in my VisiData Italian guide https://github.com/ondata/guidaVisiData/blob/master/testo/README.md#Salvare-una-tabella-HTML-in-CSV-a-partire-da-una-pagina-web

Thank you againg

[saulpw]

Fixed for html loader in f55de38; requires changes in other loaders with a sheet index.

[saulpw]

To-do to resolve this issue:

1. Fix loaders with sheet index to have rowdef sheets.
2. Write above requirement into book/loaders.md.
3. Improve startup with large files to remove sync(); file should load sync, cursor should jump after load completes (including ^C), or after sheet/row/col is available, if possible.

[anjakefala]

The IndexSheet has been developed (see visidata/sheets.py). It contains the attribute rowtype = 'sheets' on default.

Loaders to be ported:

• html
• xls
• xlsx
• xlsb
• hdf5
• sqlite
• postgres

Misc:

• requirements needs to be added to loaders.md

[saulpw]

CLI syntax is +:<sheet>:<row>:<col>.

• +:subsheet:: to ignore row/col
• can name toplevel source index if more than one: +toplevel:subsheet::

[aborruso]

Hi @saulpw if I run

curl -L "https://en.wikipedia.org/wiki/Olympic_medal" | vd -f html +:table_2:1:1

vd does not open the table_e. What's wrong in my command?

vd 2 is really great!

[anjakefala]

Hey @aborruso!

Can you please open a bug report, and link to this issue?

There is not a good way for me to remember to check up on this potential bug, otherwise. 😅