Automatic word detection

Automatic word detection is the recommended method for determining words. It determines words for all languages and no understanding of the word parsing mechanism documented in this section is required.

Rules for automatic word detection

The following rules apply:
  1. Umlauts, accented characters, and numbers are all processed as part of word.
  2. Sequence numbers if detected in columns 73–80 and columns 1–6 are ignored for data sets that are fixed with a logical record length of 80.
  3. Data set names or words joined by periods produce a long word including the periods and the individual words. For example, FMN.SFMNMOD1, produces the words FMN, SFMNMOD1 and FMN.SFMNMOD1 in the index.
  4. COBOL or words joined by dashes produce a long word including the dashes and the individual words. For example, REC-TYPE produces the words REC, TYPE and REC-TYPE in the index.
  5. A word prefixed with the ampersand character (&) produces the word with and without the prefix in the index. For example, &MYVAR produces the words &MYVAR and MYVAR in the index.