[ < < < Home ] [ < < Reference Start ] [ < Reference Contents ]
[ < Previous=PDGREPPE Replace Symbols ] [ Next=PDGREPPE Replace Controls > ]
PDGREPPE Replace Sections

Symbols like

	&	%	-	+	?

are used to replace areas of a match with assorted values.

When a match is found it encompasses two main areas, either the
entire matched area, represented by the "&" symbol, or various
subsections of the match area that have been located with the
Marked Group operators "(" and ")" in the search pattern.

So if a search pattern is

	"(bananas) or (apples)"

and the replacement pattern is

	"%1 or %0"

and a match is found, then the replacement becomes

	"apples or bananas"

the REVERSE of what is found : "bananas or apples".

This is simple swapping at work!

Or another example, if the search pattern is

	"again"

and the replacement pattern is

	"&\, and &\, and &"

and a match is found, then the replacement becomes

	"again, and again, and again".

Use of these symbols is a convenient way to reduce the number
of characters entered as a replacement, or to rearrange them in
some way with LESS typing or literal representation.

SPECIAL SYMBOL

 &		insert all of data match


	pdgreppe "TextTools" -kekikwkr"& & &" file_id.diz


..."& & &" inserts "TextTools" in a copy of file_id.diz 3
times with spaces.

SPECIAL SYMBOL FORM

 %<n>           insert <n>th Marked Group of data match


	pdgreppe "(Intelligence) Services" -kekikwkr"& %0" file_id.diz


Here:

	& %0

inserts entire contents of match:

	"&" or "Intelligence Services"

a space:

	" "

then

	"%0"

contents of Marked Group 0 or "Intelligence" to become

	Intelligence Services Intelligence

Likewise:

 %<VName>	insert Marked Group of data match identified by <VName>


	pdgreppe "(Intelligence)=intel Services" -kekikwkr"& %intel" file_id.diz


So:

	& %intel

inserts entire contents of match:

	"&" or "Intelligence Services"

a space

	" "

then

	"%intel"

to repeat the data in the Marked Group identified by "intel"
(Intelligence) to become

	Intelligence Services Intelligence

SPECIAL SYMBOL FORM

 +<n>{<pat>}	 insert pattern <pat> if Passed <n>th Marked Group

This is an UNCONDITIONAL replacement function that replaces
"<pat>" if the <n>th or <named> marked (group) is ANY OR ALL of
the match area.


	pdgreppe "Re(lease)" -kekikwkr"+0{Version}" file_id.diz


..."Re(lease)" is found and

	"+0{Version}"

has changed

	"Release"

to

	"Version"

because Marked Group 0

	(lease)

was found.

The change is made WITHOUT REGARD to the difference between
Marked Group 0's width and the entire match width of

	"Re(lease)"

The other variety,

 +<VName>{<pat>} insert pattern <pat> if Passed Marked Group
 		 identified by <VName>

works likewise:


	pdgreppe "Re(lease)=lea" -kekikwkr"+lea{Version}" file_id.diz


	"Re(lease)=lea"

is found and

	"+lea{Version}"

has changed "Release" to "Version" because the Marked Group is
named or tagged by "=lea"

	(lease)

was found.

Again, the change is made WITHOUT REGARD to the difference
between Marked Group "lea" width and the entire match width of

	"Re(lease)"

SPECIAL SYMBOL FORM

 ?<n>{<pat>}	 insert pattern <pat> if Passed <n>th Marked Group
 		 and Marked Group width is ALL of match

This is a conditional replacement function that allows a
replacement of "<pat>" to be made only if the <n>th marked
(group) is all of the match area.


	pdgreppe "(Release)" -kekikwkr"?0{Version}" file_id.diz


..."(Release)" is found and "?0{Version}" has changed "Release"
to "Version" because

	Marked Group 0 (Release) was found

and

	is ALL of match area.


	pdgreppe "Re(lease)" -kekikwkr"?0{Version}" file_id.diz


However with this pattern "Re(lease)" is found and "?0{Version}"
could DELETE "Release" rather than change it to "Version"
because

	Marked Group 0 (lease) was found

and

	is NOT ALL of match area.

So the replace pattern is empty "?0{Version}" and a delete is
done instead.

Other variety:

 ?<VName>{<pat>} insert pattern <pat> if Passed Marked Group
 		 identified by <VName>
 		 and Marked Group width is ALL of match


	pdgreppe "(Release)=rel" -kekikwkr"?rel{Version}" file_id.diz


..."(Release)=rel" is found and "?rel{Version}" has changed
"Release" to "Version" because

	Marked Group identified by (Release)=rel was found

and

	is ALL of match area.

BEWARE

A mistake can occur when a SINGLE MINUS MARK starts a replace
pattern exactly after option -kr:

It is interpreted as meaning "turn off" option -kr.  To overcome
this either start the pattern with a double quote or whitespace
between -kr and the pattern.

eg

	-kr"-meToo{XX}"

	-kr -9{XX}

SPECIAL SYMBOL FORM

 -<n>{<pat>}	 insert pattern <pat> if NOT Passed <n>th Marked Group

 -<VName>{<pat>} insert pattern <pat> if NOT Passed Marked Group
 		 identified by <VName>

These are like the +<n> and ?<n> forms just shown but replace
with the substitution pattern {<pat>} if the (Marked Group)
search pattern segment identified by <n> or <VName> FAILED.

It is good for making a selective insert if something was NOT
found in a search pattern.

eg


	pdgreppe "pattern(s)?" -kekikwkr":U&-0{isation}"  file_id.diz


finds lines in FILE_ID.DIZ with PDGREPPE and FGREPPE and
CAPITALISES with the ":U" all-to-upper-case function of the
match to Search term "pattern(s)?".

And for the line with FGREPPE where "pattern" is WITHOUT an "s"
at its end, "(s)" of "(s)?" DOES NOT MATCH because there is NO
"s" after "pattern", PDGREPPE performs an additional paste of
"isation" that also becomes upper-case to become

	PATTERNISATION

Summary of Selective Marked Group Replace functions:

The

	+<n>, ?<n> and -<n> {<pat>}

Edit Replace functions offer a way to change any of a number of
different edit patterns to be completely different by
substitution.


	pdgreppe "(Search)|(by)" -kekikwkr"?0{FIND}?1{WITH}" file_id.diz


...the replacement pattern

	"?0{Find}?1{with}"

changes the match of strings "Search" or "by" in search pattern

	(Search)|(by)

to become entirely different strings "Find" or "with" if pattern
match is made in Marked Groups 0 OR 1.


	pdgreppe "(<\l+>)|(<\u+>)" -kekwkr"?0{Peas}?1{BEANS}" file_id.diz


...changes any lower-case word

	"(<\l+>)"

or upper-case word

	"(<\u+>)"

to become entirely different strings

	"Peas" or "BEANS"

if a match is made in Marked Groups 0 or 1.

Great Stuff!  You can edit between plenty different patterns in
one swoop provided You keep track of the Marked Group alignments
or variable names.

All <n> numbers or <VName> names given for Marked Groups in Edit
Patterns BY DEFAULT MUST correspond to the Marked Groups in the
upper level MAIN User Pattern, NOT Marked Groups in any
@Definition used by the upper level user pattern or another
pattern @Definition.

However, with editing option "-kg", names given for Marked
Groups can be accessed as GLOBAL names that reference UNIQUE
Marked Groups.  Option "-jg" should also be used with this
option to make a total search pattern only contains UNIQUE
Marked Group names.

Be careful with replace patterns containing &, %<n> and
[+?-]<n>{<dat>} because they can also DELETE data if the
original match area had zero width (& and %<n>) or if the
replacement pattern data has zero width ([+?-]<n>{<empty>}).

See PDGREPPE Options "-jg" and "-kg".
[ < < < Home ] [ < < Reference Start ] [ < Reference Contents ]
[ < Previous=PDGREPPE Replace Symbols ] [ Next=PDGREPPE Replace Controls > ]