String recognition in idl

Question

I have the following strings:

F:\Sheyenne\ROI\SWIR32_subset\SWIR32_2005210_East_A.dat
F:\Sheyenne\ROI\SWIR32_subset\SWIR32_2005210_Froemke-Hoy.dat

and from each I want to extract the three variables, 1. SWIR32 2. the date and 3. the text following the date. I want to automate this process for about 200 files, so individually selecting the locations won't exactly work for me.

so I want:

variable1=SWIR32
variable2=2005210
variable3=East_A
variable4=SWIR32
variable5=2005210
variable6=Froemke-Hoy

I am going to be using these to add titles to graphs later on, but since the position of the text in each string varies I am unsure how to do this using strmid

Answer 1

I think you want to use a combination of STRPOS and STRSPLIT. Something like the following:

s = ['F:\Sheyenne\ROI\SWIR32_subset\SWIR32_2005210_East_A.dat', $
  'F:\Sheyenne\ROI\SWIR32_subset\SWIR32_2005210_Froemke-Hoy.dat']
name = STRARR(s.length)
date = name
txt = name
foreach sub, s, i do begin
  sub = STRMID(sub, 1+STRPOS(sub, '\', /REVERSE_SEARCH))
  parts = STRSPLIT(sub, '_', /EXTRACT)
  name[i] = parts[0]
  date[i] = parts[1]
  txt[i] = STRJOIN(parts[2:*], '_')
endforeach

You could also do this with a regular expression (using just STRSPLIT) but regular expressions tend to be complicated and error prone.

Hope this helps!