Replacing over multiple lines - all occurrences in file

Krivo · Jan 12, 2006#12006-01-12T20:15+00:00

Hi there,

Hope someone can help me out with this problem.

I want to delete a number of blocks of my document. The blocks I want to delete looks like this:

Code: Select all

                java.lang.ArrayIndexOutOfBoundsException: 65533
	at defacto.calcscript.Lexer.next_token(Lexer.java:677)
	at java_cup.runtime.lr_parser.scan(lr_parser.java:335)
	at java_cup.runtime.lr_parser.parse(lr_parser.java:563)
	at defacto.calcscript.Interpreter.parseScript
                (empty line)

The starting text will always be 'java.lang.xxxxx' and this text will always be the first part of the line (tabs and spaces might occur).

'java.lang.xx' can also appear other places in the document but if 'java.lang.xx' is not found on the start of a line then the text must NOT be deleted.

The end of the text part I want to delete will always be an empty line.

I tried something like this: %[^t ]++^(java.lang*^)^(*%^p) - but it didn't work

Can anyone help me out here.

thanks

Bego · Jan 13, 2006#22006-01-13T12:05+00:00

Hi Krivo,

try this one (deleteBlockRe):

Code: Select all

InsertMode
ColumnModeOff
HexOff
UnixReOn
Top
Loop 
Find RegExp "^java.lang.*"
IfFound
Find RegExp Select "^$"
IfNotFound
ExitMacro
EndIf
Cut 
Else
ExitMacro
EndIf
EndLoop

Code: Select all

BLA empty line

java.lang.ArrayIndexOutOfBoundsException: 65533
at defacto.calcscript.Lexer.next_token(Lexer.java:677)
at java_cup.runtime.lr_parser.scan(lr_parser.java:335)
at java_cup.runtime.lr_parser.parse(lr_parser.java:563)
at defacto.calcscript.Interpreter.parseScript


bla java.lang.ArrayIndex I shound not be deleted bla


java.lang.ArrayIndexOutOfBoundsException: 65533
at defacto.calcscript.Lexer.next_token(Lexer.java:677)
at java_cup.runtime.lr_parser.scan(lr_parser.java:335)
at java_cup.runtime.lr_parser.parse(lr_parser.java:563)
at defacto.calcscript.Interpreter.parseScript


other stuff

java.lang.ArrayIndexOutOfBoundsException: 65533
at defacto.calcscript.Lexer.next_token(Lexer.java:677)
at java_cup.runtime.lr_parser.scan(lr_parser.java:335)
at java_cup.runtime.lr_parser.parse(lr_parser.java:563)
at defacto.calcscript.Interpreter.parseScript

Dont forget empty line at the end ...

will look like this then:

Code: Select all

BLA empty line



bla java.lang.ArrayIndex I shound not be deleted bla




other stuff


Dont forget empty line at the end ...

If you do NOT want Unix regular expression in your environment, switch it back to off at the end.

rds Bego

Mofi · Jan 13, 2006#32006-01-13T12:14+00:00

A multi-line UltraEdit/Unix regular expression Replace All cannot be used here. But this macro will do the job.

Code: Select all

InsertMode
ColumnModeOff
HexOff
UnixReOff
Loop 
Find RegExp "%[ ^t]++java.lang"
IfNotFound
ExitLoop
EndIf
StartSelect
Find Select "^p^p"
IfSel
Key UP ARROW
Key UP ARROW
Key END
EndSelect
Delete
Else
ExitLoop
EndIf
EndLoop
UnixReOn

Remove the last UnixReOn command, if you use regular expression in UltraEdit style by default instead of Unix style.
For UltraEdit v11.10c and lower see Advanced - Configuration - Find - Unix style Regular Expressions.
For UltraEdit v11.20 and higher see Advanced - Configuration - Searching - Unix style Regular Expressions.
Macro commands UnixReOn/UnixReOff modifies this setting.

And don't forget to enable the macro property Continue if a Find with Replace not found.

This macro works for files with DOS terminated lines or Unix/Mac files which are converted automatically or after confirmation temporarily to DOS during editing. For opened Unix files in real Unix mode use ^n^n instead of ^p^p and for Mac files opened not in DOS mode use ^r^r.

Update:
Later with UE v12.00 the Perl regular expression engine was introduced in UltraEdit. With the Perl regular expression engine it is possible to do this task with a regular expression Replace All command.

The Perl regular expression search string needed for the example is: (?s)^[\t ]*java\.lang\..*?\r\n[\t ]*\r\n

Bego · Jan 13, 2006#42006-01-13T12:25+00:00

Guten Morgen Mofi,

seems we both stood up the same time

As we can see, many ways leed to Rome.
Krivo, just take the solution you like more

rds Bego

Krivo · Jan 17, 2006#52006-01-17T03:18+00:00

Hi Rego and Mofi,

Thank you very much for your input. You answer so fast - very impressive. Much appreciated!!

I ended up making a macro which looked like below. I added a gotoline so I can replace in the whole document. Further, I added this line:
Find RegExp Select "^p[ ^t]++^p" - just because I was editing the document and happened to press return at the last 'stacktrace' - and the macro would not find the indented line. But this was not part of the problem description - and will probably not be needed in reality

All the best to you guys

Krivo

Code: Select all

InsertMode
ColumnModeOff
HexOff
UnixReOff
GotoLine 1
Key HOME
Loop 
Find RegExp "%[ ^t]++java.lang"
IfNotFound
ExitLoop
EndIf
StartSelect
Find RegExp Select "^p[ ^t]++^p"
IfSel
Key UP ARROW
Key UP ARROW
Key END
EndSelect
Delete
Else
ExitLoop
EndIf
EndLoop

Apr 24, 2006#62006-04-24T15:56+00:00

Hi guys,

Just to let you know. I changed the script to something a lot more simple which is much more efficient - much faster.

Code: Select all

InsertMode
ColumnModeOff
HexOff
UnixReOff
Find RegExp "%[ ^t]++java.lang*^p^p"
Replace All ""

Take care

Krivo

Mofi · Apr 25, 2006#72006-04-25T10:48+00:00

Hi Krivo!

Yes, the *^p^p method works for selecting a block, but only because of a bug of UltraEdit. I have used this method in previous versions of my SortLanguage macro, but replaced it with the correct method in the current version.

In help of UltraEdit you can read for * : Matches any number of occurrences of any character except newline.

So your regular expression should only find a 2 line block where the first line starts with java.lang and the line below is a blank line. Well, you can use this faster replace but always take into consideration that you use a bug of UE and it maybe will not work in a future version of UE anymore when the bug is fixed.