In general I do too, but it hasn’t prevailed so far.

In the case of Wikipedia, there’s a much more straightforward possibility. The wikitable markup is essentially CSV with pipes instead of commas:

| 0:00 || [[The Kills]] || “[[Sour Cherry]]”

Why not put a bug next to every table that extracts its contents as CSV or tab-delimited?

We need (easier) ways to publish data as XML documents or fragments, or at the least to have a way to tag data on pages to make it easily consumable.

Some ideas:

- publish pages like that as XML with an XSLT stylesheet for presentation;

- modify HTML standards to allow some sort of tag, which can contain only table elements within it.

I tend to favor the XML method – if you want the page to be presented as HTML it’s fairly trivial to transform XML to HTML and CSS using XSL and we’re not heading down the direction of screwing up HTML and browser (in)compatibility.

Good grief, you’re right! How did I not know, or discover, this?

Well, it’s a sure bet I’m not alone in my ignorance. This leads to two questions:

1. What uncommon search strategy would have located this nugget?

2. How would the nugget need to have been packaged in order to yield to common search strategies?

Anyway, thanks Alf!

Probably so. Interesting to be reminded by these examples, though, what a grab-bag clipboard formats can be, and therefore how unpredictable the results.

I wonder if we’re just conditioned to look for the programmable, re-usable solution and so don’t spot that the so-simple-it-doesn’t-need-re-using solution is easier.

Results that aren’t in a HTML table require more fudging in Word: scan/replace the spaces with some other character which can then be replaced with a tab so the whole thing can then be made into a table and cut/pasted to Excel. Usually for example names are separated by a single space, where the virtual columns of data are separated by two or more spaces, so a careful scan/replace strategy can get it all lined up.

I’m going to try the Data-> from Web next, thanks ;-)

Why/how did you apply Text to Columns then?

Almost certainly. That’s why I floated this item.

> But it would be one of the last things
> that I’d try.

Fascinating, isn’t it.

When I try this from IE the CTRL-V result is as above: mostly good, some correction needed, correction complicated by active hyperlinks.

When I try from FF the CTRL-V result is garbage. But Paste Special -> As Text gives the mostly-good result without problematic hyperlinks.

@Adrian: Yeah, that’s basically what I did. Select the text from Wikipedia, paste into Excel. I had to tweak some of the cells after that, but that’s pretty much it.

I was thinking about how my non-techie girlfriend would approach the problem. Having observed her cut-and-pasting things from web pages into Word when she wants to reorganise them, I think she’d try the same with Excel.

Lo and behold, selecting all the relevant text from that wikipedia page, switching to Excel and hitting Ctrl-V gives me it all in separate cells automatically.

But it would be one of the last things that I’d try.