extract data from a string (filename)

[mnorthcott]

Hi.  

I have filenames that follow a certain pattern, but aren't always the same number of characters long.  I wondered how it is possible to extract each part of the filename.

For example:

M&N C-3456 Cutomer Name REVA 101234A1

where:

M&N = Constant company name (won't change)

C-3456 = Job Number

Customer Name = Self-Explanatory, indefinite length and number of words

**REVA = Revision Number (i would only need the A part)

**101234A1 = Main company job number (101234) and Add-on number (A1, also not always present when job number is)

** not always present

I was thinking if only wcmatch returned the exact strings it found to match the wildcard patterns I would be in business.  I'm guessing theres something obvious i'm not getting.

[Jason.Piercey]

Extract them into a list?

(parse "M&N C-3456 Cutomer Name REVA 101234A1" (chr 32))

; creates multiple strings from a single string
; [s] - string, string to breakdown
; [d] - string, separator
; return: list of strings
(defun parse (s d / i l tmp rtn)
 (cond
  ((/= "" d)
   (setq l (strlen d))
   (while (setq i (vl-string-search d s))
    (setq tmp (substr s 1 i))
    (setq rtn (cons tmp rtn))
    (setq s (substr s (1+ (+ l (strlen tmp))) (strlen s)))
    )
   (vl-remove "" (reverse (cons s rtn)))
   )
  (t s)
  )
 )

[Jason.Piercey]

You can then use SUBSTR to get portions of each string as required.

[Kent1Cooper]

As a start, Search this Forum for things like "split string" or "subdivide string", and you'll find quite a number of routines that can take a string and break it into pieces around whatever delimiting character(s) you want [in this case, the space between each piece].  I think most of them will put the separate pieces into a list of text strings, but some may do different things with them.  When they're in a list, if the pieces always come in the same order, you could get [for example] the job number from the list with (nth) or (cadr).  You could extract the A off the end of "REVA" using (substr).  What you do with the Main Company job number and add-on would depend on certain things [e.g. is the number part before the add-on always the same number of characters long?], but ways can be worked out to determine whether or not there's an add-on, and/or to separate the number from the add-on, if such things are needed.

[pbejse]

---

@mnorthcott wrote:

.....

**REVA = Revision Number (i would only need the A part)

**101234A1 = Main company job number (101234) and Add-on number (A1, also not always present when job number is)

** not always present

---

And if it is present does it always start with the word REV?

And if is present does it always start with a number? or are any chararacter to indicate that it is present?

Main company job number can be present even without the REV?