Sometimes I receive a CSV file which has a carriage return inside a cell. This is not an acceptable format to a program that will use it as input.
In order to detect if an input line is split, I determined that a bad line would not have the expected number of commas in it. Is there a bash or other common unix command line tool that would allow me to count the commas in the line? If necessary, I can write a Python or Perl program to do it, but if possible, I'd like to add a line or two to an existing bash script to cause it to fail if the comma count is wrong. Any ideas?
Strip everything but the commas, and then count number of characters left:
$ echo foo,bar,baz | tr -cd , | wc -c
2
To count the number of times a comma appears, you can use something like awk:
string=(line of input from CSV file)
echo "$string" | awk -F "," '{print NF-1}'
But this really isn't sufficient to determine whether a field has carriage returns in it. Fields can have commas inside as long as they're surrounded by quotes.
What worked for me better than the other solutions was this. If test.txt has:
foo,bar,baz
baz,foo,foobar,bar
Then cat test.txt | xargs -I % sh -c 'echo % | tr -cd , | wc -c' produces
2
3
This works very well for streaming sources, or tailing logs, etc.
In pure Bash:
while IFS=, read -ra array
do
echo "$((${#array[#]} - 1))"
done < inputfile
or
while read -r line
do
count=${line//[^,]}
echo "${#count}"
done < inputfile
Try Perl:
$ perl -ne 'print 0+#{[/,/g]},"\n"'
a
0
a,a
1
a,a,a,a,a
4
Depending on what you are trying to do with the CSV data, it may be helpful to use a wrapper script like csvquote to temporarily replace the problematic newlines (and commas) inside quoted fields, then restore them. For instance:
csvquote inputfile.csv | wc -l
and
csvquote inputfile.csv | cut -d, -f1 | csvquote -u
may be the sort of thing you're looking for. See [https://github.com/dbro/csvquote][1] for the code and more information
An example Python command you could run (since it's going to be installed on most modern shells) is:
python -c "import pathlib; print({l.count(',') for l in pathlib.Path('my_file.csv').read_text().splitlines()})"
This counts the number of commas per line, then makes a set from them (so if your lines all have the same number of commas in, you'll get a set with just that number in).
Just remove all of the carriage returns:
tr -d "\r" old_file > new_file
Related
I have a file containing a list of replacement pairs (about 100 of them) which are used by sed to replace strings in files.
The pairs go like:
old|new
tobereplaced|replacement
(stuffiwant).*(too)|\1\2
and my current code is:
cat replacement_list | while read i
do
old=$(echo "$i" | awk -F'|' '{print $1}') #due to the need for extended regex
new=$(echo "$i" | awk -F'|' '{print $2}')
sed -r "s/`echo "$old"`/`echo "$new"`/g" -i file
done
I cannot help but think that there is a more optimal way of performing the replacements. I tried turning the loop around to run through lines of the file first but that turned out to be much more expensive.
Are there any other ways of speeding up this script?
EDIT
Thanks for all the quick responses. Let me try out the various suggestions before choosing an answer.
One thing to clear up: I also need subexpressions/groups functionality. For example, one replacement I might need is:
([0-9])U|\10 #the extra brackets and escapes were required for my original code
Some details on the improvements (to be updated):
Method: processing time
Original script: 0.85s
cut instead of awk: 0.71s
anubhava's method: 0.18s
chthonicdaemon's method: 0.01s
You can use sed to produce correctly -formatted sed input:
sed -e 's/^/s|/; s/$/|g/' replacement_list | sed -r -f - file
I recently benchmarked various string replacement methods, among them a custom program, sed -e, perl -lnpe and an probably not that widely known MySQL command line utility, replace. replace being optimized for string replacements was almost an order of magnitude faster than sed. The results looked something like this (slowest first):
custom program > sed > LANG=C sed > perl > LANG=C perl > replace
If you want performance, use replace. To have it available on your system, you'll need to install some MySQL distribution, though.
From replace.c:
Replace strings in textfile
This program replaces strings in files or from stdin to stdout. It accepts a list of from-string/to-string pairs and replaces each occurrence of a from-string with the corresponding to-string. The first occurrence of a found string is matched. If there is more than one possibility for the string to replace, longer matches are preferred before shorter matches.
...
The programs make a DFA-state-machine of the strings and the speed isn't dependent on the count of replace-strings (only of the number of replaces). A line is assumed ending with \n or \0. There are no limit exept memory on length of strings.
More on sed. You can utilize multiple cores with sed, by splitting your replacements into #cpus groups and then pipe them through sed commands, something like this:
$ sed -e 's/A/B/g; ...' file.txt | \
sed -e 's/B/C/g; ...' | \
sed -e 's/C/D/g; ...' | \
sed -e 's/D/E/g; ...' > out
Also, if you use sed or perl and your system has an UTF-8 setup, then it also boosts performance to place a LANG=C in front of the commands:
$ LANG=C sed ...
You can cut down unnecessary awk invocations and use BASH to break name-value pairs:
while IFS='|' read -r old new; do
# echo "$old :: $new"
sed -i "s~$old~$new~g" file
done < replacement_list
IFS='|' will give enable read to populate name-value in 2 different shell variables old and new.
This is assuming ~ is not present in your name-value pairs. If that is not the case then feel free to use an alternate sed delimiter.
Here is what I would try:
store your sed search-replace pair in a Bash array like ;
build your sed command based on this array using parameter expansion
run command.
patterns=(
old new
tobereplaced replacement
)
pattern_count=${#patterns[*]} # number of pattern
sedArgs=() # will hold the list of sed arguments
for (( i=0 ; i<$pattern_count ; i=i+2 )); do # don't need to loop on the replacement…
search=${patterns[i]};
replace=${patterns[i+1]}; # … here we got the replacement part
sedArgs+=" -e s/$search/$replace/g"
done
sed ${sedArgs[#]} file
This result in this command:
sed -e s/old/new/g -e s/tobereplaced/replacement/g file
You can try this.
pattern=''
cat replacement_list | while read i
do
old=$(echo "$i" | awk -F'|' '{print $1}') #due to the need for extended regex
new=$(echo "$i" | awk -F'|' '{print $2}')
pattern=${pattern}"s/${old}/${new}/g;"
done
sed -r ${pattern} -i file
This will run the sed command only once on the file with all the replacements. You may also want to replace awk with cut. cut may be more optimized then awk, though I am not sure about that.
old=`echo $i | cut -d"|" -f1`
new=`echo $i | cut -d"|" -f2`
You might want to do the whole thing in awk:
awk -F\| 'NR==FNR{old[++n]=$1;new[n]=$2;next}{for(i=1;i<=n;++i)gsub(old[i],new[i])}1' replacement_list file
Build up a list of old and new words from the first file. The next ensures that the rest of the script isn't run on the first file. For the second file, loop through the list of replacements and perform them each one by one. The 1 at the end means that the line is printed.
{ cat replacement_list;echo "-End-"; cat YourFile; } | sed -n '1,/-End-/ s/$/³/;1h;1!H;$ {g
t again
:again
/^-End-³\n/ {s///;b done
}
s/^\([^|]*\)|\([^³]*\)³\(\n\)\(.*\)\1/\1|\2³\3\4\2/
t again
s/^[^³]*³\n//
t again
:done
p
}'
More for fun to code via sed. Try maybe for a time perfomance because this start only 1 sed that is recursif.
for posix sed (so --posix with GNU sed)
explaination
copy replacement list in front of file content with a delimiter (for line with ³ and for list with -End-) for an easier sed handling (hard to use \n in class character in posix sed.
place all line in buffer (add the delimiter of line for replacement list and -End- before)
if this is -End-³, remove the line and go to final print
replace each first pattern (group 1) found in text by second patttern (group 2)
if found, restart (t again)
remove first line
restart process (t again). T is needed because b does not reset the test and next t is always true.
Thanks to #miku above;
I have a 100MB file with a list of 80k replacement-strings.
I tried various combinations of sed's sequentially or parallel, but didn't see throughputs getting shorter than about a 20-hour runtime.
Instead I put my list into a sequence of scripts like "cat in | replace aold anew bold bnew cold cnew ... > out ; rm in ; mv out in".
I randomly picked 1000 replacements per file, so it all went like this:
# first, split my replace-list into manageable chunks (89 files in this case)
split -a 4 -l 1000 80kReplacePairs rep_
# next, make a 'replace' script out of each chunk
for F in rep_* ; do \
echo "create and make executable a scriptfile" ; \
echo '#!/bin/sh' > run_$F.sh ; chmod +x run_$F.sh ; \
echo "for each chunk-file line, strip line-ends," ; \
echo "then with sed, turn '{long list}' into 'cat in | {long list}' > out" ; \
cat $F | tr '\n' ' ' | sed 's/^/cat in | replace /;s/$/ > out/' >> run_$F.sh ;
echo "and append commands to switch in and out files, for next script" ; \
echo -e " && \\\\ \nrm in && mv out in\n" >> run_$F.sh ; \
done
# put all the replace-scripts in sequence into a main script
ls ./run_rep_aa* > allrun.sh
# make it executable
chmod +x allrun.sh
# run it
nohup ./allrun.sh &
.. which ran in under 5 mins, a lot less than 20 hours !
Looking back, I could have used more pairs per script, by finding how many lines would make up the limit.
xargs --show-limits </dev/null 2>&1 | grep --color=always "actually use:"
Maximum length of command we could actually use: 2090490
So just under 2MB; how many pairs would that be for my script ?
head -c 2090490 80kReplacePairs | wc -l
76923
So it seems I could have used 2 * 40000-line chunks
to expand on chthonicdaemon's solution
live demo
#! /bin/sh
# build regex from text file
REGEX_FILE=some-patch.regex.diff
# test
# set these with "export key=val"
SOME_VAR_NAME=hello
ANOTHER_VAR_NAME=world
escape_b() {
echo "$1" | sed 's,/,\\/,g'
}
regex="$(
(echo; cat "$REGEX_FILE"; echo) \
| perl -p -0 -e '
s/\n#[^\n]*/\n/g;
s/\(\(SOME_VAR_NAME\)\)/'"$(escape_b "$SOME_VAR_NAME")"'/g;
s/\(\(ANOTHER_VAR_NAME\)\)/'"$(escape_b "$ANOTHER_VAR_NAME")"'/g;
s/([^\n])\//\1\\\//g;
s/\n-([^\n]+)\n\+([^\n]*)(?:\n\/([^\n]+))?\n/s\/\1\/\2\/\3;\n/g;
'
)"
echo "regex:"; echo "$regex" # debug
exec perl -00 -p -i -e "$regex" "$#"
prefixing lines with -+/ allows empty "plus" values, and protects leading whitespace from buggy text editors
sample input: some-patch.regex.diff
# file format is similar to diff/patch
# this is a comment
# replace all "a/a" with "b/b"
-a/a
+b/b
/g
-a1|a2
+b1|b2
/sg
# this is another comment
-(a1).*(a2)
+b\1b\2b
-a\na\na
+b
-a1-((SOME_VAR_NAME))-a2
+b1-((ANOTHER_VAR_NAME))-b2
sample output
s/a\/a/b\/b/g;
s/a1|a2/b1|b2/;;
s/(a1).*(a2)/b\1b\2b/;
s/a\na\na/b/;
s/a1-hello-a2/b1-world-b2/;
this regex format is compatible with sed and perl
since miku mentioned mysql replace:
replacing fixed strings with regex is non-trivial,
since you must escape all regex chars,
but you also must handle backslash escapes ...
naive escaper:
echo '\(\n' | perl -p -e 's/([.+*?()\[\]])/\\\1/g'
\\(\n
I have this peace of code:
cat BP.csv | while read line ; do
goterm=$(awk '{print $1}') ;
name=$(awk '{print $2}') ;
grep -w "$goterm" GOEA.csv | sed "s/$goterm/pi/g" ;
done
file BP.csv has this format:
GO:0008283 cell proliferation
GO:0009405 pathogenesis
GO:0010201 response to continuous far red light stimulus by the high-irradiance response system
GO:0009641 shade avoidance
while GOEA.csv has this format:
4577 GO:0006807 0.994 2014_06_01
4577 GO:0016788 0.989 2014_06_01
4577 GO:0043169 0.977 2014_06_01
4577 GO:0043170 0.963 2014_06_01
sed doesn't work. I want to change GO:0043170 for example, to string "pi", but it gives:
sed: -e expression #1, char 12: unterminated `s' command
Why?
Thanks.
You running your awk command against no input, Try this:
cat BP.csv | while read line ; do
goterm=$(awk '{print $1}' <<< "$line") ;
name=$(awk '{print $2}' <<< "$line" ) ;
grep -w "$goterm" GOEA.csv | sed "s/$goterm/pi/g" ;
done
Let's clean up this code a bit:
while read goterm name
do
grep -w "$goterm" GOEA.csv | sed "s/$goterm/pi/g"
done < BP.cvs
The problem is that your awk statements are attempting to read in from STDIN just like your while is doing. You're reading from the same input stream.
What you want to do is to pull out the values from your line. I'm using read to do this. The read statement uses the values in $IFS to separate out the input. This is normally spaces, tabs, and newlines. The read reads each variable you put on the line, and the last value read in contains the entire rest of the line.
Thus:
while read line
reads in the entire line while:
while goterm name
will break the line as
goterm="GO:0008283"
name="cell proliferation"
One more thing. When you use grep and sed together, you probably can get away with just sed:
while read goterm name
do
sed -n "/$goterm/s/$goterm/pi/gp" GOEA.csv
done < BP.csv
The format for the sed command is:
/lines/command/parameters/
So, I'm searching for lines with $goterm in them, then I am replacing $goterm with pi. The -n means don't print out the lines as sed processes them and p means to print out the lines were the substitute is located.
By the way, csv as a file suffix means comma separated values but neither file looks like it is comma separated. Are these tabs separating each field. If so, you'll need to modify $IFS to be tabs.
I would restructure that whole thing more like this:
while read goterm restofline
do
grep -w "${goterm}" GOEA.csv | sed -e "s/${goterm}/pi/g"
done < BP.csv
No reason for the awk things, as the bash read builtin will do rudimentary field splitting for you if you give it multiple variables. Also, you aren't using name anyway, so it's not needed. cat is unnecessary as well.
Depending on your exact use case, even the grep may be unnecessary, making the inner command simply sed -ne "s/${goterm}/pi/gp" GOEA.csv. Unless your purpose for the grep -w is eliminating lines where ${goterm} is a substring of a word instead of the whole word...
For future reference, inserting a set -x above your loop in your script would show you the exact commands that are being run, so that you can compare them with your expectations.
I am writing shell script for embedded Linux in a small industrial box. I have a variable containing the text pid: 1234 and I want to strip first X characters from the line, so only 1234 stays. I have more variables I need to "clean", so I need to cut away X first characters and ${string:5} doesn't work for some reason in my system.
The only thing the box seems to have is sed.
I am trying to make the following to work:
result=$(echo "$pid" | sed 's/^.\{4\}//g')
Any ideas?
The following should work:
var="pid: 1234"
var=${var:5}
Are you sure bash is the shell executing your script?
Even the POSIX-compliant
var=${var#?????}
would be preferable to using an external process, although this requires you to hard-code the 5 in the form of a fixed-length pattern.
Here's a concise method to cut the first X characters using cut(1). This example removes the first 4 characters by cutting a substring starting with 5th character.
echo "$pid" | cut -c 5-
Use the -r option ("use extended regular expressions in the script") to sed in order to use the {n} syntax:
$ echo 'pid: 1234'| sed -r 's/^.{5}//'
1234
Cut first two characters from string:
$ string="1234567890"; echo "${string:2}"
34567890
pipe it through awk '{print substr($0,42)}' where 42 is one more than the number of characters to drop. For example:
$ echo abcde| awk '{print substr($0,2)}'
bcde
$
Chances are, you'll have cut as well. If so:
[me#home]$ echo "pid: 1234" | cut -d" " -f2
1234
Well, there have been solutions here with sed, awk, cut and using bash syntax. I just want to throw in another POSIX conform variant:
$ echo "pid: 1234" | tail -c +6
1234
-c tells tail at which byte offset to start, counting from the end of the input data, yet if the the number starts with a + sign, it is from the beginning of the input data to the end.
Another way, using cut instead of sed.
result=`echo $pid | cut -c 5-`
I found the answer in pure sed supplied by this question (admittedly, posted after this question was posted). This does exactly what you asked, solely in sed:
result=\`echo "$pid" | sed '/./ { s/pid:\ //g; }'\``
The dot in sed '/./) is whatever you want to match. Your question is exactly what I was attempting to, except in my case I wanted to match a specific line in a file and then uncomment it. In my case it was:
# Uncomment a line (edit the file in-place):
sed -i '/#\ COMMENTED_LINE_TO_MATCH/ { s/#\ //g; }' /path/to/target/file
The -i after sed is to edit the file in place (remove this switch if you want to test your matching expression prior to editing the file).
(I posted this because I wanted to do this entirely with sed as this question asked and none of the previous answered solved that problem.)
Rather than removing n characters from the start, perhaps you could just extract the digits directly. Like so...
$ echo "pid: 1234" | grep -Po "\d+"
This may be a more robust solution, and seems more intuitive.
This will do the job too:
echo "$pid"|awk '{print $2}'
How can I join multiple lines into one line, with a separator where the new-line characters were, and avoiding a trailing separator and, optionally, ignoring empty lines?
Example. Consider a text file, foo.txt, with three lines:
foo
bar
baz
The desired output is:
foo,bar,baz
The command I'm using now:
tr '\n' ',' <foo.txt |sed 's/,$//g'
Ideally it would be something like this:
cat foo.txt |join ,
What's:
the most portable, concise, readable way.
the most concise way using non-standard unix tools.
Of course I could write something, or just use an alias. But I'm interested to know the options.
Perhaps a little surprisingly, paste is a good way to do this:
paste -s -d","
This won't deal with the empty lines you mentioned. For that, pipe your text through grep, first:
grep -v '^$' | paste -s -d"," -
This sed one-line should work -
sed -e :a -e 'N;s/\n/,/;ba' file
Test:
[jaypal:~/Temp] cat file
foo
bar
baz
[jaypal:~/Temp] sed -e :a -e 'N;s/\n/,/;ba' file
foo,bar,baz
To handle empty lines, you can remove the empty lines and pipe it to the above one-liner.
sed -e '/^$/d' file | sed -e :a -e 'N;s/\n/,/;ba'
How about to use xargs?
for your case
$ cat foo.txt | sed 's/$/, /' | xargs
Be careful about the limit length of input of xargs command. (This means very long input file cannot be handled by this.)
Perl:
cat data.txt | perl -pe 'if(!eof){chomp;$_.=","}'
or yet shorter and faster, surprisingly:
cat data.txt | perl -pe 'if(!eof){s/\n/,/}'
or, if you want:
cat data.txt | perl -pe 's/\n/,/ unless eof'
Just for fun, here's an all-builtins solution
IFS=$'\n' read -r -d '' -a data < foo.txt ; ( IFS=, ; echo "${data[*]}" ; )
You can use printf instead of echo if the trailing newline is a problem.
This works by setting IFS, the delimiters that read will split on, to just newline and not other whitespace, then telling read to not stop reading until it reaches a nul, instead of the newline it usually uses, and to add each item read into the array (-a) data. Then, in a subshell so as not to clobber the IFS of the interactive shell, we set IFS to , and expand the array with *, which delimits each item in the array with the first character in IFS
I needed to accomplish something similar, printing a comma-separated list of fields from a file, and was happy with piping STDOUT to xargs and ruby, like so:
cat data.txt | cut -f 16 -d ' ' | grep -o "\d\+" | xargs ruby -e "puts ARGV.join(', ')"
I had a log file where some data was broken into multiple lines. When this occurred, the last character of the first line was the semi-colon (;). I joined these lines by using the following commands:
for LINE in 'cat $FILE | tr -s " " "|"'
do
if [ $(echo $LINE | egrep ";$") ]
then
echo "$LINE\c" | tr -s "|" " " >> $MYFILE
else
echo "$LINE" | tr -s "|" " " >> $MYFILE
fi
done
The result is a file where lines that were split in the log file were one line in my new file.
Simple way to join the lines with space in-place using ex (also ignoring blank lines), use:
ex +%j -cwq foo.txt
If you want to print the results to the standard output, try:
ex +%j +%p -scq! foo.txt
To join lines without spaces, use +%j! instead of +%j.
To use different delimiter, it's a bit more tricky:
ex +"g/^$/d" +"%s/\n/_/e" +%p -scq! foo.txt
where g/^$/d (or v/\S/d) removes blank lines and s/\n/_/ is substitution which basically works the same as using sed, but for all lines (%). When parsing is done, print the buffer (%p). And finally -cq! executing vi q! command, which basically quits without saving (-s is to silence the output).
Please note that ex is equivalent to vi -e.
This method is quite portable as most of the Linux/Unix are shipped with ex/vi by default. And it's more compatible than using sed where in-place parameter (-i) is not standard extension and utility it-self is more stream oriented, therefore it's not so portable.
POSIX shell:
( set -- $(cat foo.txt) ; IFS=+ ; printf '%s\n' "$*" )
My answer is:
awk '{printf "%s", ","$0}' foo.txt
printf is enough. We don't need -F"\n" to change field separator.
I have input.txt
1
2
3
4
5
I need to get such output.txt
1,2,3,4,5
How to do it?
Try this:
tr '\n' ',' < input.txt > output.txt
With sed, you could use:
sed -e 'H;${x;s/\n/,/g;s/^,//;p;};d'
The H appends the pattern space to the hold space (saving the current line in the hold space). The ${...} surrounds actions that apply to the last line only. Those actions are: x swap hold and pattern space; s/\n/,/g substitute embedded newlines with commas; s/^,// delete the leading comma (there's a newline at the start of the hold space); and p print. The d deletes the pattern space - no printing.
You could also use, therefore:
sed -n -e 'H;${x;s/\n/,/g;s/^,//;p;}'
The -n suppresses default printing so the final d is no longer needed.
This solution assumes that the CRLF line endings are the local native line ending (so you are working on DOS) and that sed will therefore generate the local native line ending in the print operation. If you have DOS-format input but want Unix-format (LF only) output, then you have to work a bit harder - but you also need to stipulate this explicitly in the question.
It worked OK for me on MacOS X 10.6.5 with the numbers 1..5, and 1..50, and 1..5000 (23,893 characters in the single line of output); I'm not sure that I'd want to push it any harder than that.
In response to #Jonathan's comment to #eumiro's answer:
tr -s '\r\n' ',' < input.txt | sed -e 's/,$/\n/' > output.txt
tr and sed used be very good but when it comes to file parsing and regex you can't beat perl
(Not sure why people think that sed and tr are closer to shell than perl... )
perl -pe 's/\n/$1,/' your_file
if you want pure shell to do it then look at string matching
${string/#substring/replacement}
Use paste command. Here is using pipes:
echo "1\n2\n3\n4\n5" | paste -s -d, /dev/stdin
Here is using a file:
echo "1\n2\n3\n4\n5" > /tmp/input.txt
paste -s -d, /tmp/input.txt
Per man pages the s concatenates all lines and d allows to define the delimiter character.
Awk versions:
awk '{printf("%s,",$0)}' input.txt
awk 'BEGIN{ORS=","} {print $0}' input.txt
Output - 1,2,3,4,5,
Since you asked for 1,2,3,4,5, as compared to 1,2,3,4,5, (note the comma after 5, most of the solutions above also include the trailing comma), here are two more versions with Awk (with wc and sed) to get rid of the last comma:
i='input.txt'; awk -v c=$(wc -l $i | cut -d' ' -f1) '{printf("%s",$0);if(NR<c){printf(",")}}' $i
awk '{printf("%s,",$0)}' input.txt | sed 's/,\s*$//'
printf "1\n2\n3" | tr '\n' ','
if you want to output that to a file just do
printf "1\n2\n3" | tr '\n' ',' > myFile
if you have the content in a file do
cat myInput.txt | tr '\n' ',' > myOutput.txt
python version:
python -c 'import sys; print(",".join(sys.stdin.read().splitlines()))'
Doesn't have the trailing comma problem (because join works that way), and splitlines splits data on native line endings (and removes them).
cat input.txt | sed -e 's|$|,|' | xargs -i echo "{}"