Member Avatar for iamthesgt

For a shell script used to automatically generate c++ code files, I have to split certain names apart for them to be formatted. I split them at a capital letter, but I do not want to split if there are multiple capital letters in sequence. For example:
I want this (and I get this):
ChanNameWordWrapMode -> Chan Name Word Wrap Mode
I want this:
ChanNameVAlign -> Chan Name VAlign
OSDSettings -> OSD Settings
But I get this:
ChanNameVAlign -> Chan Name V Align
OSDSettings -> O S D Settings.
The command I am using right now is:

`echo $string | sed 's/\([A-Z]\)/ \1/g' | sed 's/^ *//;s/ *$//'`

Can someone help with a sed command that does what I am asking? Thanks.

  • 2 Contributors
  • 6 Replies
  • 2K Views
  • 3 Days Discussion Span
  • Latest Post Latest Post by iamthesgt

Dani AI

Generated

Nice thread. nailed the main rule: split on a lowercase-to-uppercase boundary. The missing piece is handling acronym runs without breaking them up, while still splitting before the final cap that starts a normal word (so OSDSettings -> OSD Settings, but VAlign stays intact).

You can do both with two substitutions in one sed run. Use POSIX classes for portability:

# BSD/macOS: -E ; GNU sed: -r or -E
sed -E 's/([[:lower:]])([[:upper:]])/\1 \2/g; s/([[:upper:]]{2,})([[:upper:]][[:lower:]])/\1 \2/g'

Why it works:

  • First s/// inserts a space between a lowercase and an uppercase (ChanName -> Chan Name).
  • Second s/// inserts a space only after a run of 2+ capitals when the next token looks like Cap+lowercase. That gives OSD Settings, but it does not touch VAlign because there is no place where 2+ capitals are immediately followed by Cap+lowercase inside VAlign.

If you already have inputs like "O S D Settings" (as mentioned later), normalize them first by collapsing spaces between adjacent single-letter capitals, then apply the splitter:

# collapse spaced acronyms: "O S D Settings" -> "OSD Settings"
sed -E ':a; s/\b([A-Z])\s+([A-Z])\b/\1\2/g; ta'

Tips:

  • Prefer [[:upper:]]/[[:lower:]] over A-Z/a-z if locale or non-ASCII might appear.
  • To stay strictly ASCII for code identifiers, run with LC_ALL=C.

Recommended Answers

All 6 Replies

Member Avatar for nezachem

In the ChanNameVAlign case you want to split at the transition from lowercase to uppercase:
s/\([a-z]\)\([A-Z]\)/\1 \2/g

In the OSDSettings, I don't see a non-contradictory criteria.

Member Avatar for iamthesgt

Thanks. I will try that, but the current regex does split OSDSettings into O S D Settings

Member Avatar for iamthesgt

Okay, I thought of something to try, but I can't seem to construct the regex. If I have a string of the format "O S D Settings" that I want to convert to "OSD Settings", I could just concatenate the capital letters that are followed by a space, right? How might I construct this Regex?

Member Avatar for iamthesgt

This is working part way:

sed 's/ \([A-Z]\) \([A-Za-z]*\)/\1/g'

However, when run on the string "O S D Settings", it returns "OS Settings" - it is only operating on the first instance of the regex. Any tips on how to make it work for the whole line?

Member Avatar for nezachem

concatenate the capital letters that are followed by a space

I can think only of a 3-step solution (which looks ugly, but possibly it can be improved):
- replace all wanted spaces with, say, underscores:
s/ \([A-Z][^ ]\)/_\1/g'
- remove all remaining spaces:
s/ //g
- restore wanted spaces:
s/_/ /g'

Something like that...

Member Avatar for iamthesgt

Well, you don't have to use three commands - you can use pipes to redirect the code:

echo $string | sed 's/ \([A-Z][^ ]\)/_\1/g' | sed 's/ //g' | sed 's/_/ /g'

Anyway, thanks. This works. Thanks for all the help.

Be a part of the DaniWeb community

We're a friendly, industry-focused community of developers, IT pros, digital marketers, and technology enthusiasts meeting, networking, learning, and sharing knowledge.