SUBLIME with GREP: Search for words, ignoring certain lines

Question

I have a long text document that has over 24,000 lines. I need to search through for 83 different words and remove all instances of these 83 words when the words occur alone on a line (see example below). The Words will be removed. But the search must ignore:

• all lines that only have capital text
• All lines that begin with a § symbol (regardless if the rest of the line has capital or lower case letters).

Example

§History of TIME
HISTORY OF TIME
of
Future
Past
Of
Many Of
Official
Officer
North of the town
I am one of them 

If I search for Of then in the above only the third and sixth line would be highlighted.

The first line would not be highlighted because it begins with a §.

The second line would not be highlighted because it is in all caps.

Many of, North of the town, and I am one of them would not be highlighted, because it is not Of on its own.

Is this possible in Sublime? I think searching for all 83 separate words would make the query too complicated. But if I can search for one word at a time (ensuring it skips all lines that are in Caps and all lines that begin with a §, then that would be okay).

If this is not possible in Sublime, is there any other method I can use?

Answer 1

You say you want to remove all occurrences of of when a line does not start with a § or is ALLCAPS and when not part of a many of, north of, one of, etc. phrase.

Use

Find: (^(?:(?:§.*|[^[:alpha:]\n\r]*[[:upper:]]+(?:[^[:alpha:]\n\r]+[[:upper:]]+)*[^[:alpha:]\n\r]*))$|(?i:\b(?:many|north|one)\s+of\b))|(?i:\bof\b)
Replace: $1

See regex demo

Details

• (^(?:(?:§.*|[^[:alpha:]\n\r]*[[:upper:]]+(?:[^[:alpha:]\n\r]+[[:upper:]]+)*[^[:alpha:]\n\r]*))$|(?i:\b(?:many|north|one)\s+of\b)) - Capturing group 1 (referred to with $1 placeholder from the replacement pattern):
  • ^ - start of a line
  • (?: - a group:
    • §.*
    • | - or
    • [^[:alpha:]\n\r]*[[:upper:]]+(?:[^[:alpha:]\n\r]+[[:upper:]]+)*[^[:alpha:]\n\r]* - 0+ chars other than LF/CR symbols and letters, then 1+ uppercase letters, then 0+ sequences of non-linebreak and non-letter chars followed with 1+ uppercase letters and ending with 0+ sequences of non-linebreak and non-letter chars
    • )$ - end
  • | or
    • (?i:\b(?:many|north|one)\s+of\b) - a case insensitive group matching many, north or one followed with 1+ whitespaces and then of as whole words
• | - or
• (?i:\bof\b) - a case insensitive group: of as a whole word.

See the SublimeText3 test, mind to select regex mode and case sensitive matching: