Very Computer

A simple question, please help!

In a shell program, how can I change the
first character of a string to upper case
if it is in lower case?

     Ex:    unix -> Unix
            Foo  -> Foo    

Thanks a lot!

-Hongliang

0.  Warning.  It is possible that I have fail to see an obvious
and easy solution.  Proceed at your own risk.

1. Using nawk or gawk:

newstr=`echo $oldstr | nawk '{ printf "%s%s", toupper(substr($0,1,1)), substr($0,2) }'`

Comment: Not all systems have nawk or gawk

2. With expr and tr:

newstr=`expr $oldstr : '\(.\).*' | tr a-z A-Z``expr $oldstr : '.\(.*\)'`

Comment: This is quite wasteful in the number of processes it spawns

3. With sed:

newstr=`echo $oldstr | sed -f scrptfle`

where the file "scrptfle" contains the following sed script:

h
s/.\(.*\)/\1/
x
s/\(.\).*/\1/
y/abcefghijklmnopqrstuvwxyz/ABCEFGHIJKLMNOPQRSTUVWXYZ/
G
s/\n//

Comment 1: This is ugly, but it's fast and efficient.
Comment 2: The y operator is an undocumented feature of sed, so it
may not be available on all versions of sed.

4. With perl

Solution left as an expercise :-)

: In a shell program, how can I change the
: first character of a string to upper case
: if it is in lower case?
:  
:      Ex:    unix -> Unix
:             Foo  -> Foo    

This will capitalize a specific word:

sed 's/\<unix/Unix/g' filname

The \< construct tells sed to match the beginning of a word.

Hope that helps some.

--
/* tdk  --  Science Computing and Research Support (SCARS)  --  STScI */
/* Disclaimer:  The above remarks are my My MY personal observations. */

| The \< construct tells sed to match the beginning of a word.

Not all versions of sed honor '\<' and '\>'.  If any given regexp magic isn't
supported by ed and isn't in your manual page for sed, sed very likely might
not support it, even if egrep or ex does.

There's an old saying that if sed were based on ex rather than ed it would
have been called sex.

David W. Tamkin   Box 59297   Northtown Station, Illinois  60659-0297