How to write a search pattern to include a space in findstr?
Question
I want to search all files in a certain directory for occurrences of statements such as
Load frmXYZ
I am on Windows 7, using the findstr
command. I tried:
findstr /n Load.*frm *.*
But this gives me unwanted results such as:
If ABCFormLoaded Then Unload frmPQR
So I tried to put a blankspace between Load
and frm
and gave the command like this:
findstr /n Load frm *.*
But this simply searched for all occurrences of the word load
or all occurrences of the word frm
. How do I get around this problem?
Solution
If you use spaces, you need the /C:
option to pass the the literal string(s) to the regex /R
option.
Once the it gets to the regex, it's treated as a regex.
That said, this is typical MS trash.
Use two regex search strings
The bottom line is that you have to use 2 strings to handle cases where
Load frm
is at the beginning like so:
Load frm apples bananas carrots
OR in the middle like so:
some other text Load frm and more
.
Version without character classes
Below is using XP sp3, windows 7 may be different, both are trash!
findstr /N /R /C:" *Load *frm" /C:"^Load *frm" test.txt
7:Load frm is ok
8: Load frm is ok
Mind the Colon
NOTE: The colon in /C:
is MANDATORY for this to work.
If you leave out the colon then findstr
's error handling is just to treat /C
as an invalid option, ignore that invalid option and go ahead anyway. Leading to unexpected and unwanted output.
Equivalent version using character classes
findstr /N /R /C:"[ ][ ]*Load[ ][ ]*frm" /C:"^Load[ ][ ]*frm" test.txt
Character classes breakdown
// The first regex search string breaks down like this:
[ ] // require 1 space
[ ]* // optional many spaces
Load // literal 'Load'
[ ] // require 1 space
[ ]* // optional many spaces
frm // literal 'frm'
// The second regex search string breaks down like this:
^ // beginning of line
Load // literal 'Load'
[ ] // require 1 space
[ ]* // optional many spaces
frm // literal 'frm'
A real regex might be \bLoad\s+frm
OTHER TIPS
Use the /c
option:
findstr /n /c:"Load frm" *.*
From the help (findstr /?
):
/C:string Uses specified string as a literal search string.
Use word delimiter regex
I used the the special \<
"beginning of word" regex symbol.
I tried this on the Win10 version of findstr. But according to Microsoft this special \<
symbol has been in findstr.exe
ever since WinXP.
Full (and painful) breakdown of many options that do NOT work below.
At the very bottom: what actually worked.
The sample file itself
C:\>type lines.txt
Load frmXYZ // This line should match.
If ABCFormLoaded Then Unload frmPQR // This line should NOT match.
pears Load frm grapes pineapples // This line should match.
// This blank line should NOT match.
LOAD FRMXYZ // This line should match.
IF ABCFORMLOADED THEN UNLOAD FRMPQR // This line should NOT match.
PEARS LOAD FRM GRAPES PINEAPPLES // This line should match.
// This blank line should NOT match.
load frmxyz // This line should match.
if abcformloaded then unload frmpqr // This line should NOT match.
pears load frm grapes pineapples // This line should match.
Wrong. With regular execution space is treated as delimiter.
C:\>type lines.txt | findstr /N "Load frm"
1:Load frmXYZ // This line should match.
2:If ABCFormLoaded Then Unload frmPQR // This line should NOT match.
3:pears Load frm grapes pineapples // This line should match.
9:load frmxyz // This line should match.
10:if abcformloaded then unload frmpqr // This line should NOT match.
11:pears load frm grapes pineapples // This line should match.
Wrong: With Regex option space is STILL treated as delimiter.
C:\>type lines.txt | findstr /N /R "Load frm"
1:Load frmXYZ // This line should match.
2:If ABCFormLoaded Then Unload frmPQR // This line should NOT match.
3:pears Load frm grapes pineapples // This line should match.
9:load frmxyz // This line should match.
10:if abcformloaded then unload frmpqr // This line should NOT match.
11:pears load frm grapes pineapples // This line should match.
More right but still wrong. With /C option we now get preserved spaces but don't find other character cases.
C:\>type lines.txt | findstr /N /R /C:"Load frm"
1:Load frmXYZ // This line should match.
3:pears Load frm grapes pineapples // This line should match.
Wrong. /I for "Ignore Case" does not help. We get matches from within words we did not want.
C:\>type lines.txt | findstr /N /R /I /C:"Load frm"
1:Load frmXYZ // This line should match.
2:If ABCFormLoaded Then Unload frmPQR // This line should NOT match.
3:pears Load frm grapes pineapples // This line should match.
5:LOAD FRMXYZ // This line should match.
6:IF ABCFORMLOADED THEN UNLOAD FRMPQR // This line should NOT match.
7:PEARS LOAD FRM GRAPES PINEAPPLES // This line should match.
9:load frmxyz // This line should match.
10:if abcformloaded then unload frmpqr // This line should NOT match.
11:pears load frm grapes pineapples // This line should match.
Right. Use special "Beginning of word" regex symbol. Matches beginning-of-line or space.
Either case sensitive:
C:\>type lines.txt | findstr /N /R /C:"\<Load frm"
1:Load frmXYZ // This line should match.
3:pears Load frm grapes pineapples // This line should match.
or ignoring case
C:\>type lines.txt | findstr /N /R /I /C:"\<Load frm"
1:Load frmXYZ // This line should match.
3:pears Load frm grapes pineapples // This line should match.
5:LOAD FRMXYZ // This line should match.
7:PEARS LOAD FRM GRAPES PINEAPPLES // This line should match.
9:load frmxyz // This line should match.
11:pears load frm grapes pineapples // This line should match.
This piece of code will only allow letters, numbers, underscore and white space in keyword:
set /p keyword="Enter keyword: " || Set keyword=
set keyword_replaced=%keyword: =_%
echo %keyword_replaced%| findstr /r "[^0-9a-zA-Z_]" > nul
if errorlevel 1 goto noexit
echo special characters in keyword not allowed (except space and _), TERMINATING
timeout 4
exit /b 0
:noexit