Joining truncated lines in a file

Question

I have a file with broken lines as follows,

C12321 net12415431 net41432143 +
 1.079879E-17
C12322 net2135 net648641 +
 3.4659E-17

I want to create a file from this file which has lines as follows

C12321 net12415431 net41432143 1.079879E-17
C12322 net2135 net648641 3.4659E-17

I need to do this using sed or awk with in a couple of lines.

regards, Asif

Ed Morton · Answer 1 · 2013-05-10T22:39:00.390

1

sed is an excellent tool for simple substitutions on a single line but for any other text manipulation just use awk - the solution will be clearer, more easily extensible, more robust, more portable, and in many cases briefer (not that that matters much).

Read the file and remove all occurrences of " +\n" (space-plus-newline):

$ cat file
C12321 net12415431 net41432143 +
 1.079879E-17
C12322 net2135 net648641 +
 3.4659E-17

$ awk -v RS= '{gsub(/ \+\n/,""); print}' file
C12321 net12415431 net41432143 1.079879E-17
C12322 net2135 net648641 3.4659E-17

There are many alternatives, some briefer, some using less memory, etc. this is just the clearest IMHO.

edited May 10 '13 at 22:39

answered May 10 '13 at 22:25

Ed Morton

188,023
17
78
185

I would argue that if "briefer" equals "equally or more readable" than it does matter. – Adrian Frühwirth May 11 '13 at 15:19
Correct but that's the argument for conciseness, which is a desirable attribute of software, not brevity which is not. – Ed Morton May 11 '13 at 15:53

score 0 · Answer 2 · edited May 23 '17 at 12:13

0

You can do it like this:

$ sed ':a;N;$!ba;s/+\n/ /g' your_file

it will replace all + followed by a new line with empty string.

In case you want the file to be updated with this replacement, add the -i parameter:

$ sed -i ':a;N;$!ba;s/+\n/ /g' your_file

The explanation of the command can be found in https://stackoverflow.com/a/1252191/1983854 :

Update: explanation.

create a label via :a

append the current and next line to the pattern space via N

if we are before the last line, branch to the created label $!ba ($! means not to do it on the last line (as there should be one final newline)).

finally the substitution replaces every newline with a space on the pattern space (which is the whole file).

edited May 23 '17 at 12:13

Community

1
1

answered May 10 '13 at 13:37

fedorqui

275,237
103
548
598

Thank you very much. Where can I get the explanation for the command you typed? – asif iqbal May 10 '13 at 13:44
I was about to explain it when you deleted the question. Now I see it is "alive" again :) See my updated answer. – fedorqui May 10 '13 at 13:46
The command is actually not changing the file at all. All the + followed by nextline are as they were before. – asif iqbal May 10 '13 at 13:57
OK, if you need to update the file, add the `-i` flag as I just indicated in the answer. – fedorqui May 10 '13 at 13:58
I think I need to go through the sed funde before i try to understand this explanation. – asif iqbal May 10 '13 at 14:03
The main important part is that if you do `sed 's/hello/bye/g'` all "hello" will be replaced by "bye" in a file. The first part is more high-level sed so it can wait. – fedorqui May 10 '13 at 14:06
@ fedorqui The file I need to convert ha the following information. – asif iqbal May 10 '13 at 14:12
`.SUBCKT slew_ldo_core VDD_IO OUT VSS_IO TEST_SLEW_LDO_PINOUT_IO +TB_IO<6> + TB_IO<5> SS_IN PD_LDO_VDD_IO ISINK_10U FB_BACK vfb TM_OUT FS_OUT TB_IO<7> + fb_bias_5u REF_INT_500mV vim TB_IO<2> TB_IO<1> TB_IO<3> TB_IO<4> + TB_IO<0> CAP_NODE` – asif iqbal May 10 '13 at 14:12
Please update your question with this info. It is not possible to read it properly in the comments. – fedorqui May 10 '13 at 14:13
I executed sed ':a;N;$!ba;s/+\n/ /g' in terminal after invoking bash. – asif iqbal May 10 '13 at 14:19
It did not change any thing and displaied it as it is on STDIO. – asif iqbal May 10 '13 at 14:19
This content was actually in a file – asif iqbal May 10 '13 at 14:20

chepner · Answer 3 · 2013-05-12T15:24:17.337

0

A longer but more readable alternative to a single sed command:

tr '+' '\\' < tmp.txt | while read line; do printf "%s\n" "$line"; done

This just replaces + with a line continuation character that the shell built-in read understands.

edited May 12 '13 at 15:24

answered May 10 '13 at 14:03

chepner

497,756
71
530
681

This only replaced + with nothing. It didn't join the truncated lines in to single line. By the way what king of line truncation is +\n. – asif iqbal May 10 '13 at 14:25
`tr '+' '\' < tmp.txt` should simply replace each '+' with a literal backslash. The `read` command (without the `-r` option usually recommended) sees the literal "\" and continues reading until the next unescaped newline. – chepner May 10 '13 at 15:05
1

don't you need to escape the backslash in the tr so it doesn't look like escape-tick? You also need to add the variable REPLY to the read command. It's also worth mentioning that by not setting `IFS=` as usually recommended you're stripping off leading white space which in this case is desirable. – Ed Morton May 10 '13 at 22:44
I didn't realize `REPLY` as a default was a `bash`-specific extension to `read`. As for the backslash, *everything* is treated literally inside single quotes, so no escaping is necessary (or possible). That's a good point about `IFS`. – chepner May 11 '13 at 00:30
bash on cygwin: `echo "a+b" | tr '+' '\'` outputs `tr: warning: an unescaped backslash at end of string is not portable` and then `a\b` while `echo "a+b" | tr '+' '\\'` outputs `a\b`. – Ed Morton May 11 '13 at 15:58
Ah, OK. It seems different versions of `tr` will handle an unescaped backslash different. – chepner May 12 '13 at 15:07

score 0 · Answer 4 · answered May 10 '13 at 22:05

0

This might work for you (GNU sed):

sed '$!N;s/ +\n//;P;D' file

answered May 10 '13 at 22:05

potong

55,640
6
51
83

Joining truncated lines in a file

4 Answers4