Print Section of file - comp.lang.awk

Hi! How do I print a section of a file, where the section is specified by shell variables. Something like awk -v fi="$1" fu="$2" '/^fi/,/^fu/{print}' file does not work. This also keeps true for appending the shell variables, i.e. awk '...' fi="$1" fu="$2". However, when I substitute fi, fu with the desired values, the output is as desired. With thanks Yours Wolfgang

On Wednesday 8 April 2009 20:02, Gehricht@googlemail.com wrote: > Hi! > > How do I print a section of a file, where the section is specified by > shell variables. Something like > > awk -v fi="$1" fu="$2" '/^fi/,/^fu/{print}' file try awk -v fi="$1" -v fu="$2" '$0~f1,$0~fu' file

On Wednesday 8 April 2009 20:06, pk wrote: > On Wednesday 8 April 2009 20:02, Gehricht@googlemail.com wrote: > >> Hi! >> >> How do I print a section of a file, where the section is specified by >> shell variables. Something like >> >> awk -v fi="$1" fu="$2" '/^fi/,/^fu/{print}' file > > try > > awk -v fi="$1" -v fu="$2" '$0~f1,$0~fu' file oops, you need to match against "^"fi and "^"fu.

On Apr 8, 1:02=A0pm, "Gehri...@googlemail.com" <Gehri...@googlemail.com> wrote: > Hi! > > How do I print a section of a file, where the section is specified by > shell variables. Something like > > awk -v fi=3D"$1" fu=3D"$2" '/^fi/,/^fu/{print}' file > > does not work. This also keeps true for appending the shell variables, > i.e. awk '...' fi=3D"$1" fu=3D"$2". However, when I substitute fi, fu wit= h > the desired values, the output is as desired. > With thanks > Yours Wolfgang Are fi and fu line numbers or regular expressions or something else? Depending on the answer, one of these MAY be what you want: awk -v fi=3D"$1" -v fu=3D"$2" 'NR>=3Dfi; NR=3D=3Dfu{exit}' file awk -v fi=3D"$1" -v fu=3D"$2" '$0~"^"fi , $0~"^"fu"' file Ed.

Thanks (to both of you) for your suggestions! > > awk -v fi="$1" -v fu="$2" '$0~"^"fi , $0~"^"fu' file > Yes, that is exactly, what I was looking for. Just for the sake of curiousity, could you explain me, how this works? With thanks Yours Wolfgang

On Apr 9, 3:21=A0am, "Gehri...@googlemail.com" <Gehri...@googlemail.com> wrote: > Thanks (to both of you) for your suggestions! > > > > > awk -v fi=3D"$1" -v fu=3D"$2" '$0~"^"fi , $0~"^"fu' file > > Yes, that is exactly, what I was looking for. Just for the sake of > curiousity, could you explain me, how this works? > With thanks > Yours Wolfgang Does anyone else find it strange that on this file f1 aa baa baaa baaa ccc ccc ccc ddd awk '/^aa/,/^b/' f1 awk '$0~"^aa",$0~"^b"' f1 and awk '$0~"^""aa",$0~"^""b"' f1 all yield the expected output aa baa but the similar awk '"^aa","^b"' f1 awk '"^aa","b"' f1 and awk '"^aa",/b/' f1 all yield a different output, aa baa baaa baaa ccc ccc ccc ddd And one might also try awk '/aa/,"b"' f1 awk '/aa/,/b/' f1 and awk '/^aa/,"aa"' f1 to be really puzzled!

On Friday 10 April 2009 01:02, r.p.loui@gmail.com wrote: > Does anyone else find it strange that on this file f1 > > aa > baa > baaa > baaa > ccc > ccc > ccc > ddd > > awk '/^aa/,/^b/' f1 > awk '$0~"^aa",$0~"^b"' f1 > and > awk '$0~"^""aa",$0~"^""b"' f1 > > all yield the expected output > > aa > baa > > > but the similar > > awk '"^aa","^b"' f1 > awk '"^aa","b"' f1 > and > awk '"^aa",/b/' f1 > > all yield a different output, > > aa > baa > baaa > baaa > ccc > ccc > ccc > ddd > > And one might also try > > awk '/aa/,"b"' f1 > awk '/aa/,/b/' f1 > and > awk '/^aa/,"aa"' f1 > > to be really puzzled! That is simply expected behavior. Using a string alone just evaluates to true (unless it's the empty string), so these programs awk '"^aa","^b"' f1 awk '"^aa","b"' f1 awk '"^aa",/b/' f1 are equivalent to awk '1,1' f1 awk '1,1' f1 awk '1,/b/' f1 and the output is as expected (if you understand how ranges work). Similarly, these awk '/aa/,"b"' f1 awk '/aa/,/b/' f1 awk '/^aa/,"aa"' f1 are equivalent to awk '/aa/,1' f1 awk '/aa/,/b/' f1 awk '/^aa/,1' f1 and the results are again as expected.

r.p.loui@gmail.com wrote: > On Apr 9, 3:21 am, "Gehri...@googlemail.com" <Gehri...@googlemail.com> > wrote: > >>Thanks (to both of you) for your suggestions! >> >> >> >> >>>awk -v fi="$1" -v fu="$2" '$0~"^"fi , $0~"^"fu' file >> >>Yes, that is exactly, what I was looking for. Just for the sake of >>curiousity, could you explain me, how this works? >>With thanks >>Yours Wolfgang > > > Does anyone else find it strange No. > that on this file f1 > > aa > baa > baaa > baaa > ccc > ccc > ccc > ddd > > awk '/^aa/,/^b/' f1 > awk '$0~"^aa",$0~"^b"' f1 > and > awk '$0~"^""aa",$0~"^""b"' f1 > > all yield the expected output > > aa > baa > > > but the similar > > awk '"^aa","^b"' f1 > awk '"^aa","b"' f1 > and > awk '"^aa",/b/' f1 > > all yield a different output, > > aa > baa > baaa > baaa > ccc > ccc > ccc > ddd > > And one might also try > > awk '/aa/,"b"' f1 > awk '/aa/,/b/' f1 > and > awk '/^aa/,"aa"' f1 > > to be really puzzled! All output as expected. ("string" is no regexp but a predicate.) Janis

On Apr 9, 7:44=A0pm, pk <p...@pk.invalid> wrote: > On Friday 10 April 2009 01:02, r.p.l...@gmail.com wrote: > > > > > > > Does anyone else find it strange that on this file f1 > > > aa > > baa > > baaa > > baaa > > ccc > > ccc > > ccc > > ddd > > > =A0awk '/^aa/,/^b/' f1 > > =A0awk '$0~"^aa",$0~"^b"' f1 > > and > > =A0awk '$0~"^""aa",$0~"^""b"' f1 > > > all yield the expected output > > > aa > > baa > > > but the similar > > > =A0 awk '"^aa","^b"' f1 > > =A0 awk '"^aa","b"' f1 > > and > > =A0 awk '"^aa",/b/' f1 > > > all yield a different output, > > > aa > > baa > > baaa > > baaa > > ccc > > ccc > > ccc > > ddd > > > And one might also try > > > =A0 awk '/aa/,"b"' f1 > > =A0 awk '/aa/,/b/' f1 > > and > > =A0 awk '/^aa/,"aa"' f1 > > > to be really puzzled! > > That is simply expected behavior. Using a string alone just evaluates to > true (unless it's the empty string), so these programs > > awk '"^aa","^b"' f1 > awk '"^aa","b"' f1 > awk '"^aa",/b/' f1 > > are equivalent to > > awk '1,1' f1 > awk '1,1' f1 > awk '1,/b/' f1 > > and the output is as expected (if you understand how ranges work). > > Similarly, these > > awk '/aa/,"b"' f1 > awk '/aa/,/b/' f1 > awk '/^aa/,"aa"' f1 > > are equivalent to > > awk '/aa/,1' f1 > awk '/aa/,/b/' f1 > awk '/^aa/,1' f1 > > and the results are again as expected.- Hide quoted text - > > - Show quoted text - Ah, I did not know that. Thanks. So why can I compose regexps from strings in other contexts where a regexp is expected, such as FS=3D"a""b" or match(x,"^"string) and so on?

On Friday 10 April 2009 18:03, r.p.loui@gmail.com wrote: > So why can I compose regexps from strings in other contexts where a > regexp is expected, > such as > > FS="a""b" > > or > > match(x,"^"string) Because if awk expects a regex, whatever it finds is interpreted as a regex. Specifically, FS *can* be a regex (if it's not a single character, then awk assumes it's a regex). And of course, match() expects its second argument to be a regex. Of course, all that does not mean that a string is *always* treated as a regex (in fact it never is, unless the context requires a regex).

On Apr 10, 12:23=A0pm, pk <p...@pk.invalid> wrote: > On Friday 10 April 2009 18:03, r.p.l...@gmail.com wrote: > > > So why can I compose regexps from strings in other contexts where a > > regexp is expected, > > such as > > > FS=3D"a""b" > > > or > > > match(x,"^"string) > > Because if awk expects a regex, whatever it finds is interpreted as a reg= ex. > Specifically, FS *can* be a regex (if it's not a single character, then a= wk > assumes it's a regex). And of course, match() expects its second argument > to be a regex. > Of course, all that does not mean that a string is *always* treated as a > regex (in fact it never is, unless the context requires a regex). So a range a,b does not require a pair of regexp's a and b? Right. It can be a relational expression or a boolean combination of patterns... So the non-null string in that context evaluates to 1 (null string to 0) because awk chooses first to see it as a relational expression rather than a regular expression? Seems a bit unfortunate..., but I do like the rule you formulated -- "in fact it never is, unless the context requires a regex"... is this documented in Robbins?

r.p.loui@gmail.com wrote: > On Apr 10, 12:23 pm, pk <p...@pk.invalid> wrote: > >>On Friday 10 April 2009 18:03, r.p.l...@gmail.com wrote: >> >> >>>So why can I compose regexps from strings in other contexts where a >>>regexp is expected, >>>such as >> >>>FS="a""b" >> >>>or >> >>>match(x,"^"string) >> >>Because if awk expects a regex, whatever it finds is interpreted as a regex. >>Specifically, FS *can* be a regex (if it's not a single character, then awk >>assumes it's a regex). And of course, match() expects its second argument >>to be a regex. >>Of course, all that does not mean that a string is *always* treated as a >>regex (in fact it never is, unless the context requires a regex). > > > So a range a,b does not require a pair of regexp's a and b? Right. Yes. > It can be > a relational expression or a boolean combination of patterns... Not quite. It's just a _predicate_. (Some like to call it condition.) (With the special case extension of a range defined by predicates as in the examples of this thread.) Or consider relational expressions which actually evaluate again to a _predicate_. See also thread "What is an awk statement ?" (2009-03-17) in c.l.a. > So > the non-null string > in that context evaluates to 1 (null string to 0) because awk chooses > first to see it as > a relational expression rather than a regular expression? > > Seems a bit unfortunate..., The unfortunate thing is the long established misnomer "pattern" in pattern { action } instead of predicate { action } /pattern/ { action } where in the latter case /pattern/ is a shortcut for $0 ~ /pattern/ which is actually, again, a predicate. Janis > but I do like the rule you formulated -- > "in fact it never is, unless the context requires a regex"... is this > documented in Robbins?

On Friday 10 April 2009 22:21, r.p.loui@gmail.com wrote: >> Because if awk expects a regex, whatever it finds is interpreted as a >> regex. Specifically, FS *can* be a regex (if it's not a single character, >> then awk assumes it's a regex). And of course, match() expects its second >> argument to be a regex. >> Of course, all that does not mean that a string is *always* treated as a >> regex (in fact it never is, unless the context requires a regex). > > So a range a,b does not require a pair of regexp's a and b? Right. > It can be > a relational expression or a boolean combination of patterns... So > the non-null string > in that context evaluates to 1 (null string to 0) because awk chooses > first to see it as > a relational expression rather than a regular expression? Right. Ranges are fairly generic, so you can do awk 'NR==4,NR==10' to print from line 4 to 10, or awk 'NR==4,0' to print from line 4 to the end, or awk '$5<6,$5>20' to print from a line where 5th field is lesser than 6 to a line where it's greater than 20 (assuming that makes sense for the particular problem). It may be worth noting that when you do something like awk '/foo/,/bar/' you are effectively doing awk '$0~/foo/,$0~/bar/' due to the way regex literals behavior is defined in that context. Plain strings are NOT defined to behave that way anywhere. > Seems a bit unfortunate..., but I do like the rule you formulated -- > "in fact it never is, unless the context requires a regex"... is this > documented in Robbins? It's documented (implicitly, possibly explicity - I didn't check) in the awk language specification. I suggest you read the specification for awk. http://www.opengroup.org/onlinepubs/9699919799/utilities/awk.html

On Apr 10, 5:02=A0pm, pk <p...@pk.invalid> wrote: > On Friday 10 April 2009 22:21, r.p.l...@gmail.com wrote: > > >> Because if awk expects a regex, whatever it finds is interpreted as a > >> regex. Specifically, FS *can* be a regex (if it's not a single charact= er, > >> then awk assumes it's a regex). And of course, match() expects its sec= ond > >> argument to be a regex. > >> Of course, all that does not mean that a string is *always* treated as= a > >> regex (in fact it never is, unless the context requires a regex). > > > So a range a,b does not require a pair of regexp's a and b? =A0Right. > > It can be > > a relational expression or a boolean combination of patterns... =A0So > > the non-null string > > in that context evaluates to 1 (null string to 0) because awk chooses > > first to see it as > > a relational expression rather than a regular expression? > > Right. Ranges are fairly generic, so you can do > > awk 'NR=3D=3D4,NR=3D=3D10' to print from line 4 to 10, or > > awk 'NR=3D=3D4,0' to print from line 4 to the end, or > > awk '$5<6,$5>20' to print from a line where 5th field is lesser than 6 to= a > line where it's greater than 20 (assuming that makes sense for the > particular problem). > > It may be worth noting that when you do something like > > awk '/foo/,/bar/' > > you are effectively doing > > awk '$0~/foo/,$0~/bar/' > > due to the way regex literals behavior is defined in that context. Plain > strings are NOT defined to behave that way anywhere. > > > Seems a bit unfortunate..., but I do like the rule you formulated -- > > "in fact it never is, unless the context requires a regex"... is this > > documented in Robbins? > > It's documented (implicitly, possibly explicity - I didn't check) in the = awk > language specification. I suggest you read the specification for awk. > > http://www.opengroup.org/onlinepubs/9699919799/utilities/awk.html Aha, that is nicer reading than the man pages for sure. Here is the relevant passage: Patterns A pattern is any valid expression, a range specified by two expressions separated by a comma, or one of the two special patterns BEGIN or END. Expression Patterns An expression pattern shall be evaluated as if it were an expression in a Boolean context. If the result is true, the pattern shall be considered to match, and the associated action (if any) shall be executed. If the result is false, the action shall not be executed. which would make the rule explicit in this document (it is not clear from the man page, I contend). Since strings are presumably "valid expressions" and "expression patterns", they get evaluated in a Boolean context (as a predicate, hence, true for non-null and false for null). I would even add something like "Regular expressions are also valid expression patterns. Unlike other contexts in which a string can be treated as a regular expression, strings cannot function like regular expressions in the pattern or range context." That would be helpful for people who are expecting the coercion from string to regexp...

On Tuesday 14 April 2009 19:00, r.p.loui@gmail.com wrote: > Patterns > A pattern is any valid expression, a range specified by two > expressions separated by a comma, or one of the two special patterns > BEGIN or END. > > Expression Patterns > An expression pattern shall be evaluated as if it were an > expression in a Boolean context. If the result is true, the pattern > shall be considered to match, and the associated action (if any) shall > be executed. If the result is false, the action shall not be executed. > > which would make the rule explicit in this document (it is not clear > from the man page, I contend). Since strings are presumably "valid > expressions" and "expression patterns", they get evaluated in a > Boolean context (as a predicate, hence, true for non-null and false > for null). I would even add something like "Regular expressions are > also valid expression patterns. Unlike other contexts in which a > string can be treated as a regular expression, strings cannot function > like regular expressions in the pattern or range context." It's the other way round. This is the relevant part: "When an ERE token appears as an expression in any context other than as the right-hand of the '˜' or "!˜" operator or as one of the built-in function arguments described below, the value of the resulting expression shall be the equivalent of: $0 ˜ /ere/" *That* is the special case, and it's clearly (imho) understandable as such. Knowing how that is evaluated by awk, you can then read the part you quoted (even if it comes first in the page). It's not written anywhere (nor it can be inferred) that the above would or should be expected to happen for "normal" data types like strings and numbers. > That would be helpful for people who are expecting the coercion from > string to regexp... As I said, from what's written in the language description, one should not expect that to happen.

Similiar Articles:

Printing CHM file cuts off text - comp.text.pdf
When opening the chm file the only thing different is that i can see horizontal scroll bar on the sections that are not printing right. So i have to scroll to the right ...

PDF Printout Garbage - comp.text.pdf
Using Adobe Acrobat 9.3.1 on a Vista Home Premium 64-bit dual-core system. When I view a PDF file on my screen, it looks fine, but when I go to print...

Script-Fu: ERROR: unbound variable - comp.graphics.apps.gimp ...
But then instead of opening the image I would like to save it to the file. Before this I need to flatten the image or at least make the text being present in the main ...

goto in expect scripts - comp.lang.tcl
Set PDF file as background in layout - comp.databases.filemaker ... Print Section of file - comp.lang.awk Can I set up a PDF to have selectable sections print via a check ...

how to save or print a filemaker Pro 6.0 job as .pdf - comp ...
If you're using MacOS X, then you can use FileMaker's normal Print function to "print" to a PDF file. ... you might even find it at the same website in the Windows section ...

sed: printing only between {} - comp.unix.shell
Unix Sed Tutorial: Printing File Lines using Address and Patterns Printing Operation in Sed. Linux Sed command allows you to print only specific lines based on the line ...

problem with pcl duplex printing - comp.os.linux.misc
I have strange problem when i print ps file in duplex by command lpr my priter works ... None: " -dDuplex=false" *CloseUI: *Duplex ISTR that I have edited this section ...

how to change the background color and print the PDF file? - comp ...
Hi all, I have a PDF file which was converted from Powerpoint file. The background was unfornately blue and the foreground color is white. If I print it out directly ...

convert integer to string - comp.lang.perl.misc
Yesterday, my user requested the app to print a calculated field, which is a ... 3D> @courses, xlist =3D> '', } } close FAC; open SEC, '<', "$sections_file", or ...

delete columns csv file - comp.lang.awk
... out = out sep $i sep=FS } print out }' file file Regards ... sorry to re-start with this issue of reading > the file twice, would you direct me towards the section ...

Export as eps with embedded bitmap - comp.soft-sys.matlab ...
If i just use the print name -depsc2 I end up with a rather large file that doesn't convert well to a pdf. ... Ideally I'd like to export this so that the colour section is ...

printer not printing multiple copies - comp.sys.sun.admin ...
It's available if you search www.hp.com's download section for the words ... like remedy 1 when you print a Postscript file, but not when you print text files or PCL ...

extract parts of file - newbie - comp.lang.perl.misc
Limit selection to rows 5000-7000 Check out $. in perlvar and read the section ... The last 5 lines of the entire file print @LINES[ $#LINES-4 .. $#LINES ...

Set PrimoPDF on Windows to DEFAULT to PRINT resolution - comp.text ...
how to change the background color and print the PDF file? - comp ... Set PDF file as ... to PRINT resolution ... Can I set up a PDF to have selectable sections print via a ...

can u suggest me to do better than this? - comp.lang.awk ...
... lines in one file named s then stored the 'r' lines in one file named r i.e awk '/^-/ {print $12 ... to zero (or null) so you don't need the above BEGIN section. > ARGIND ...

sed: print section of file from string to end of file
Hi, I've got a file with garbage (empty lines, weird chars) before the real content of my file. In this case, I am deleting everything before the '<

Printing excel files : A Handy guide
These sections have a common header which I want to repeat for parts of section I would like to print. Yes we will print different sections of an Excel File.

7/24/2012 4:49:51 AM