awk without printing newline

问题:

I want the variable sum/NR to be printed side-by-side in each iteration. How do we avoid awk from printing newline in each iteration ? In my code a newline is printed by default in each iteration

for file in cg_c ep_c is_c tau xhpl
printf "\n $file" >> to-plot.xls
    for f in 2.54 1.60 800 
        awk '{sum+=$3}; END  {print  sum/NR}' ${file}_${f}_v1.xls >> to-plot-p.xls
    done
done

I want the output to appear like this

cg_c ans1  ans2  ans3  
ep_c ans1  ans2  ans3 
is_c ans1  ans2  ans3
tau  ans1  ans2  ans3 
xhpl ans1  ans2  ans3

my current out put is like this

**cg_c**
ans1
ans2
ans3
**ep_c**
ans1
ans2
ans3
**is_c**
ans1
ans2
ans3
**tau**
ans1
ans2
ans3
**xhpl**
ans1
ans2
ans3

回答1:

awk '{sum+=$3}; END {printf "%f",sum/NR}' ${file}_${f}_v1.xls >> to-plot-p.xls

print will insert a newline by default. You dont want that to happen, hence use printf instead.

回答2:

The ORS (output record separator) variable in AWK defaults to "\n" and is printed after every line. You can change it to " " in the BEGIN section if you want everything printed consecutively.

回答3:

_{I guess many people are entering in this question looking for a way to avoid the new line in awk. Thus, I am going to offer a solution to just that, since the answer to the specific context was already solved!}

In awk, print automatically inserts a ORS after printing. ORS stands for "output record separator" and defaults to the new line. So whenever you say print "hi" awk prints "hi" + new line.

This can be changed in two different ways: using an empty ORS or using printf.

Using an empty ORS

awk -v ORS= '1' <<< "hello
man"

This returns "helloman", all together.

The problem here is that not all awks accept setting an empty ORS, so you probably have to set another record separator.

awk -v ORS="-" '{print ...}' file

For example:

awk -v ORS="-" '1' <<< "hello
man"

Returns "hello-man-".

Using printf (preferable)

While print attaches ORS after the record, printf does not. Thus, printf "hello" just prints "hello", nothing else.

$ awk 'BEGIN{print "hello"; print "bye"}'
hello
bye
$ awk 'BEGIN{printf "hello"; printf "bye"}'
hellobye

Finally, note that in general this misses a final new line, so that the shell prompt will be in the same line as the last line of the output. To clean this, use END {print ""} so a new line will be printed after all the processing.

$ seq 5 | awk '{printf "%s", $0}'
12345$
#    ^ prompt here

$ seq 5 | awk '{printf "%s", $0} END {print ""}'
12345

回答4:

one way

awk '/^\*\*/{gsub("*","");printf "\n"$0" ";next}{printf $0" "}' to-plot.xls

回答5:

If Perl is an option, here is a solution using fedorqui's example:

seq 5 | perl -ne 'chomp; print "$_ "; END{print "\n"}'

Explanation:
chomp removes the newline
print "$_ " prints each line, appending a space
the END{} block is used to print a newline

output: 1 2 3 4 5