Re: Old Timers ??

["J.D. Smith" <jdsmith(at)astro.cornell.edu>]

Eric Kihn wrote:
> 
> Hello all,
> 
>     I've go some legacy code I'm trying to update and it looks like:
> 
>     ; Read text header from file
>     ; until 'end header' reached.
>     ;-----
>     line = ''
>     header = ''
>     repeat begin
>         readf, unit, line
>         if (n_elements(header) eq 0) then $
>         begin
>             header = line
>         endif else $
>         begin
>             header = [header,line]
>         endelse
>     endrep until (strtrim(line,2) eq 'end header')
> 
>     ;-----
>     ; Close file.
>     ;-----
>     close, unit & free_lun, unit
> 
>     ;-----
>     ;Get relevant info from the header.
>     ;-----
>     dda_header.file_id = $
>         strtrim( get_keyword( header, 'file ID'), 2)
> 
> This code is circa 95 and it's balking at the get_keyword() function.  So my
> questions are two:
> 1) Did previous versions of IDL have a get_keyword() (I'm 5.3 now)?
> 2) It's clear this code is simulating a HASH with Arrays, does IDL 5.3 have
> native hash support of some kind?

1) No, but a search at
http://www.astro.washington.edu/deutsch/idl/htmlhelp/index.html reveals:
;+
; Project     : SOHO - CDS
;
; Name        : GET_KEYWORD
;
; Purpose     : Extract values in a string array that appear after
;               keyword construct such as: KEYWORD=VALUE
;               (e.g. extract all time values following
STARTIME=time_value)
;

Could be from that, but then again, it might be something else
altogether.  

2) No, sadly.  The get_keyword routine probably does a linear search
through the header line array, looking for "file ID" in this case.  This
is of course just the type of thing hashes were made for.  The
get_keyword routine I found just uses "where", as you might expect. 
This means not only is the search linear, all elements are considered,
even after a match is found.  It could have been made more efficient in
IDL 5.4 if array_equal() had been implemented not just to return true as
soon as a match is found, but return the index of that match.  Oh well. 
Maybe an optional variable will be added to array_equal for this in
5.4.1.  (hint)

It is possible to make a better simulation of a hash with a structure,
but that is somewhat inflexible since adding new keywords (fields) is
difficult, without copying the whole thing, though create_struct() can
do it.  Then there is the matter of indexing the structure with a
variable field (the essence of what hashes do).  Possible, but not
efficient... in fact nearly as inefficient as a "where" on the full
array.  Unfortunately, it's not that easy to write a good hash type that
scales well for small and large hashes, and doesn't eat too much
memory.  

JD


-- 
 J.D. Smith                  |   WORK: (607) 255-6263
 Cornell Dept. of Astronomy  |         (607) 255-5842
 304 Space Sciences Bldg.    |    FAX: (607) 255-5875  
 Ithaca, NY 14853            |