如何从head命令到sed添加到文件的第一行？

Question

I would like to add the text which is the result from head command to sed . Can I achieve this in a one liner?

This is what I have come up so far.

header=`head -n 1 csv1.csv`
sed -e '1 i\'$'\n''$header' another_csv.csv

What I don't know is how do I pass the result from one command to another command. I tried usnig xargs but no luck. I don't know how to pass that as a variable.

Answer 1

除非我缺少任何东西，否则可能是：

head -1 csv1.csv && cat another.csv

Answer 2

This might work for you (GNU sed):

sed -e '1r file2' -e 'q' file1

or perhaps use:

head -1 file1 | cat - file2

Answer 3

awk :

awk 'NR==1{getline h<"csv1.csv";print h}7 ' another.csv

test:

kent$  head f c
==> f <==
1
2
3
4
5

==> c <==
10
9
8

kent$  awk 'NR==1{getline h<"c";print h}7 ' f
10
1
2
3
4
5

gnu sed

kent$  sed '1s/.*/head -n 1 c;echo &/ge' f 
10
1
2
3
4
5

Answer 4

@anubhava's solution is simpler, but just to show the corrected version of the OP's approach:

Linux

sed -e "1 i $(head -n 1 csv1.csv)" another_csv.csv

• On Linux you need not place the text passed to i on a separate line (thus no need for \\ and the literal newline ( $'\\n' ).
• In the original approach, '$header' was placed in single quotes, preventing string interpolation - double quotes are needed: "$header" (in my reformulation I used double quotes for the entire sed command (not recommended in general) and directly placed the command inside, using command substitution ( $(...) ).

OSX

Same as above, except \\ and a newline ( $'\\n' ) must be inserted right after the i :

sed -e '1 i\'$'\n'"$(head -n 1 csv1.csv)" another_csv.csv

(The equivalent of:

sed -e "1 i\\
$(head -n 1 csv1.csv)
" another_csv.csv

)