multicolumn ARRAY SORT

[Erich Schulman]

I will be reading a CSV file with 15 columns and x (normally about 500) rows. It will be hard to tell visually whether I am getting the results I expect, so I wanted to come here first. The entire array will be DIMed starting from 1,1 rather than 0,0.

I have to sort first on column 3, then sort on column 12, and finally sort on column 2. I do not want any columns re-arranged. The data in column 5 will stay in column 5, but the entire row would move up or down in the array as the columns require.

How should I set up these ARRAY SORTs to be sure I get the intended result?

[Michael Mattias]

I would just build a sort key using a single dimension array...

TYPE SORTKEY
   Col3    AS STRING * whatever
   Col12  AS STRING  * whatever
   Col2    AS STRING  * whatever 
END TYPE

  LOCAL SK () AS SORTKEY , CSV() AS STRING

  hFile = FreeFile 
  OPEN csvFile FOR INPUT AS hFile 
  FILESCAN hfile, RECORDS TO nRec 
  REDIM   CSV(1 TO nrec)     ' I would use zero based personally 
  REDIM   SK (1 to Rec) 
  LINE   INPUT #Hfile, CSV() 
  CLOSE hFile 
  FOR Z = 1 TO nRec 
     SK(Z).Col3 = PARSE$(CSV(Z),3) 
     SK(Z).Col12 = PARSE$(CSV(Z),12) 
     SK(Z).Col2= PARSE$(CSV(Z),2) 
  NEXT
  ARRAY SORT SK(), TAGARRAY CSV()

CSV() is now in order as requested. Do what you will with it.

MCM