[REBOL] Little data trick

From: jeff::rebol::com at: 18-May-2001 11:09

  Here's a little trick.

  Say you've got a little database you're making and sometimes
  you want to use only one field and sometimes you want to use
  two fields.  For example you want to have a block of email
  and urls, but sometimes you only want to do searches over
  the emails and sometimes you want to get the urls from the
  emails as a key.

  Someone might do this:

email-url-db: [
    [a--example--com] http://www.example.com/a
    [b--example--net] http://www.example.net/b
    [c--example--org] http://www.example.org/c
    [d--example--edu] http://www.example.edu/d
]
just-emails: copy []
foreach [em url] email-url-db [append just-emails em]
just-emails: make hash! just-emails
email-url-db: make hash! email-url-db

  That's okay, but you can be sneaky if you save your database
  on disk so that it looks like this:

;=====  %example-db.dat  =================

[a--example--com]  ;  http://www.example.com/a
[b--example--net]  ;  http://www.example.net/b
[c--example--org]  ;  http://www.example.org/c
[d--example--edu]  ;  http://www.example.edu/d

;=========================================

  Now you can just do:

dbf: %example-db.dat
just-emails: load dbf
email-url-db: load replace/all read dbf ";" #

  So when your script gets a new email and url you just add to the
  database:

write/append dbf reform [new-email ";" new-url newline]

  You can of course extend the principal to as many fields as
  you like, put you just have to strip as many needed
  semicolons from each line before loading (read/lines).

  Important Caveat: You have to be sure that semicolon won't
  show up legitimately in any of your fields (safe with emails
  and urls).  So while saving you would check to make sure the
  data doesn't contain the separator and change it
  appropriately.  Anyhow, I just thought the trick was cute
  and figured I'd share. YMMV.

  -jeff

  [This is the only message in the thread]