CODEWITHSUNDEEP
parsingrebol
Rebnoob
Rebnoob

Reputation: 133

Best way to parse this in Rebol

How do I extract the transaction receipt datetime with the least bit of noise in my parse rule from the following HTML? (The output I'm looking to get is this: "Transaction Receipt: 04/28/2011 17:03:09")

 <FONT COLOR=DARKBLUE>Transaction Receipt </FONT></TH></TR><TR></TR><TR></TR><TR><TD COLSPAN=4 ALIGN=CENTER><FONT SIZE=-1 COLOR=DARKBLUE>04/28/2011 17:03:09</FONT>

The following works but I don't get a good feeling! There is guaranteed to be a datetime following the words Transaction Receipt somewhere (although I wouldn't do a greedy match if I'm doing a grep)

parse d [
    thru {<FONT COLOR=DARKBLUE>Transaction Receipt </FONT></TH></TR><TR></TR><TR></TR><TR><TD COLSPAN=4 ALIGN=CENTER><FONT SIZE=-1 COLOR=DARKBLUE>}
    copy t to "</FONT>"
    ]

Upvotes: 0

Views: 265

Answers (2)

rgchris
rgchris

Reputation: 3718

Timely as per usual: if the format is consistent, you can always try to explicitly match dates:

rule: use [dg tag date value][
    tag: use [chars][
        chars: charset [#"a" - #"z" #"A" - #"Z" #"0" - #"9" " =-"]
        ["<" opt "/" some chars ">"]
    ]

    date: use [dg mo dy yr tm][
        dg: charset "0123456789"
        [
            copy mo [2 dg "/"] copy dy [2 dg "/"] copy yr 4 dg
            " " copy tm [2 dg ":" 2 dg ":" 2 dg]
            (value: load rejoin [dy mo yr "/" tm])
        ]
    ]

    [
        some [
              "Transaction Receipt" (probe "Transaction Receipt")
            | date (probe value)

            ; everything else
            | some " " | tag ; | skip ; will parse the whole doc...
        ]
    ]
]

Upvotes: 1

Carl Read
Carl Read

Reputation: 11

This is shorter...

parse d [thru <FONT SIZE=-1 COLOR=DARKBLUE> copy t to </FONT>]

but isn't specifically looking for the datetime pair. And unfortunately REBOL considers the date used an invalid one...

>> 04/28/2011
** Syntax Error: Invalid date -- 04/28/2011
** Near: (line 1) 04/28/2011

so you can't search for it specifically. If the date was 28/04/2011 (and there was a space after the time, though why it's needed for load I'm not sure), the following would work...

parse load d [to date! copy t to </FONT>]

Hmmm. Try this...

t: ""
parse d [
    some [
        to "<" thru ">" mark: copy text to "<" (if text [append t text]) :mark
    ]
]

That returns: "Transaction Receipt 04/28/2011 17:03:09"

It works by skipping all the tags, appending any text that's left to t.

Hope that helps!

Upvotes: 1

Related Questions