sed or awk on flat file with no separators swap data on line

[Ratzie]

I have a flat file the is 773 charecters long.

What I need to do is swap two data fields around.
Both are 38 characters, so I do not need to pad. (blank spaces)

I need to swap char 1-38 data field with char 87-124 data field for each line.

I loop will do, but will I need to store these in variables?
Appreciate the help.

[Mel Burslan]

Not elegant but simple solution:

f1=`cat myfile | cut -c 1-38`
f2=`cat myfile | cut -c 39-86`
f3=`cat myfile | cut -c 87-124`
f4=`cat myfile | cut -c 125-`

echo ${f3}${f2}${f1}${f4} > mynewfile

[H.Merijn Brand (procura]

Is perl an option?

# perl -pe'@f=unpack"a38a48a3a*",$_;$_=join"",@f[2,1,0,3]' file

Many other options available. TMTOWTDI

Enjoy, Have FUN! H.Merijn

[James R. Ferguson]

Hi:

# perl -nle 'print $3.$2.$1.$4 if m/(.{38})(.{48})(.{38})(.{649})/' filename

Regards!

...JRF...

[Jonathan Fife]

cat file | sed -e '/\(......................................\)\(....................................................\)\(......................................\)\(.*\)/\3\2\1\4/g'

[Ratzie]

Here is a twist...
Thanks for the perl works great, how about only changing the lines that at the 206 & 207 char = XG?

Thanks

[James R. Ferguson · Accepted solution]

Hi (again):

# perl -nle 'if (m/.{205}XG/ and m/(.{38})(.{48})(.{38})(.{649})/) {print $3.$2.$1.$4} else {print}' filename

Regards!

...JRF...

[Ratzie]

Awesome expect on the else, if adds a ^M at the end to increase the line to 774.

[James R. Ferguson]

Hi (again):

Perhaps:

# perl -ne 'chomp;if (m/.{205}XG/ and m/(.{38})(.{48})(.{38})(.{649})/) {print $3.$2.$1.$4} else {print}' filename

Regards!

...JRF...

[H.Merijn Brand (procura]

Not that I disapprove of the regex solutions, note however that pack/unpack is both easier and about 23% faster:

# perl -MBenchmark=:all -le'$a="a"x800;cmpthese(-2,{re=>q{$a=~m/(.{38})(.{48})(.{38})(.{649})/},pu=>q{unpack"a38a48a3a*",$a}})'
Rate re pu
re 263161/s -- -19%
pu 323570/s 23% --

worse for the .{238} thing, which is wrong anyway, as it ix not anchored. Please use substr there:

# perl -MBenchmark=:all -le'$a="aXG"x300;cmpthese(-2,{re=>q{$a=~m/^.{205}XG/},pu=>q{(unpack"x205a2",$a)[0]eq"XG"},ss=>q{substr($a,205,2)eq"XG"}})'
Rate re pu ss
re 782184/s -- -10% -66%
pu 868848/s 11% -- -63%
ss 2327702/s 198% 168% --

Regular expressions are powerful, and fun, but not the right tool for the given problem

Enjoy, Have FUN! H.Merijn

[Ratzie]

So how would I take what you did and combine the two for use?

[Hein van den Heuvel]

It's hidden in the benchmark code;

# perl -pe'if (substr($a,205,2)eq"XG") {@f=unpack"a38a48a3a*",$_;$_=join"",@f[2,1,0,3]}' file

>> flat file with no separators

seperators are overrated and a waste of space, cpu time, and flexibility.


The extra carriage return may have been an artifact of the subsequent transfer to OpenVMS

Please consider looking arounf for the right tools on OpenVMS to do the job 'in place'.

That field in position 205 may well have been an 'alternate key' and so you might have been able to jump straight to those target records withou reading any others... on OpenVMS.