merge lines based on specific last pattern of a line

ymho · #1 (**permalink**) 01-07-09, 07:23

pls advise how to merge the following lines (by sed command)

1;
2
3
4
5;
6
7;

to becomes
1
2345
67

vgersh99 · #2 (**permalink**) 01-07-09, 10:55

Code:

nawk -v RS=';' '$1=$1' myFile

ymho · #3 (**permalink**) 01-07-09, 11:29

thx vgersh99 ... how about

1;
2
3
4
;5;
6
7;

expected result:1) 2) no data changed except lines merging 2) only last ; of each merged line is replaced
=>
1
234;5
67

vgersh99 · #4 (**permalink**) 01-07-09, 11:51

Code:

nawk '{a=(a=="")? $0 : a OFS $0}; /;$/ { print substr(a, 1, length(a)-1);a=""}' myFile

ymho · #5 (**permalink**) 01-07-09, 12:02

woo... but donno why my result contains space between fields ... pls advise
os : sun unix (soloris)

mine:
1
2 3 4 ;5
6 7

vgersh99 · #6 (**permalink**) 01-07-09, 12:10

Code:

nawk '{a=(a=="")? $0 : a OFS $0}; /;$/ { gsub(OFS, "", a);print substr(a, 1, length(a)-1);a=""}' myFile

ymho · #7 (**permalink**) 01-08-09, 05:53

sorry vgersh99, i mislead you .. my problem is merge the broken line due to content in text file:
e.g
1st data line <EOL>
2nd data
line <EOL>
3rd
data
line <EOL>
4th data line <EOL>
=>
1st data line <EOL>
2nd data line <EOL>
3rd data line <EOL>
4th data line <EOL>

pls advise

vgersh99 · #8 (**permalink**) 01-08-09, 11:12

Given the previous solution, what have you tried and where exactly are you stuck?

ymho · #9 (**permalink**) 01-08-09, 11:43

$cat tmp3
1st data line <EOL>
2nd data
line <EOL>
3rd
data
line <EOL>
4th data line <EOL>

command:
nawk '{a=(a=="")? $0 : a OFS $0}; /;$/ { gsub(OFS, "", a);print substr(a, 1, length(a)-1);a=""}' tmp3
=>
no result

command:
nawk '{a=(a=="")? $0 : a OFS $0}; />$/ { gsub(OFS, "", a);print substr(a, 1, length(a)-1);a=""}' tmp3
=>
1stdataline<EOL
2nddataline<EOL
3rddataline<EOL
4thdataline<EOL

mike_bike_kite · #10 (**permalink**) 01-08-09, 11:43

This might work :

Code:

cat your_file | gawk '{
                split( $0, a, "" )
                for( v in a ) {
                        if ( a[v] >= 0 && a[v] <=9 ) printf a[v]
                        if ( a[v] == ";" ) print "\n"
                }
        }' | grep "."

But it assumes you only want to print out the chars 0-9.

Mike

vgersh99 · #11 (**permalink**) 01-08-09, 11:58

Code:

nawk -v sep='<EOL>' '{a=(a=="")? $0 : a OFS $0}; $NF ~ (sep "$") { print substr(a, 1, length(a) - length(sep));a=""}' myFile

ymho · #12 (**permalink**) 01-08-09, 12:02

oh but my unix version has no gawk command... btw, can sed work ? checked from sed command, it can allow us to merge line selectively but donno why to code.

ymho · #13 (**permalink**) 01-08-09, 12:09

Quote:

Originally Posted by vgersh99

Code:

nawk -v sep='<EOL>' '{a=(a=="")? $0 : a OFS $0}; $NF ~ (sep "$") { print substr(a, 1, length(a) - length(sep));a=""}' myFile

revised as following in order to keep <EOL>

nawk -v sep='<EOL>' '{a=(a=="")? $0 : a OFS $0}; $NF ~ (sep "$") { print a;a=""}' tmp3
=>
1st data line <EOL>
2nd data line <EOL>
3rd data line <EOL>
4th data line <EOL>
=>
extra spaces found b/w "2nd data " & "3rd..." line

vgersh99 · #14 (**permalink**) 01-08-09, 12:13

not sure what you're implying and/or what to change.
If you're looking to get rid of the extra spaces, look into the previous solution (hint: use 'gsub').

In the future, please use the vB Codes when posting data or code samples.

mike_bike_kite · #15 (**permalink**) 01-08-09, 14:24

gawk and awk are very similar, both are usually available with gawk has more features. I've rewritten the code extensively to use awk :

Code:

cat your_file | awk '{
                split( $0, a, "" )
                for( v in a ) {
                        if ( a[v] >= 0 && a[v] <=9 ) printf a[v]
                        if ( a[v] == ";" ) print "\n"
                }
        }' | grep "."

Mike

Internet Services

Web Development

Digital Marketing

Consumer Tech