如何将一个文本文件分割为多个 * . txt 文件？

小开

最佳答案

You can use the Linux Bash core utility split:

split -b 1M -d  file.txt file

Note that M or MB both are OK but size is different. MB is 1000 * 1000, M is 1024^2

If you want to separate by lines you can use -l parameter.

UPDATE

a=(`wc -l yourfile`) ; lines=`echo $(($a/12)) | bc -l` ; split -l $lines -d  file.txt file

Another solution as suggested by Kirill, you can do something like the following

split -n l/12 file.txt

Note that is l not one, split -n has a few options, like N, k/N, l/k/N, r/N, r/k/N.

小开

Using Bash:

readarray -t lines < file.txt
count=${#lines[@]}


for i in "${!lines[@]}"; do
index=$(( (i * 12 - 1) / count + 1 ))
echo "${lines[i]}" >> "file${index}.txt"
done

Using AWK:

awk '{
a[NR] = $0
}
END {
for (i = 1; i in a; ++i) {
x = (i * 12 - 1) / NR + 1
sub(/\..*$/, "", x)
print a[i] > "file" x ".txt"
}
}' file.txt

Unlike split, this one makes sure that the number of lines are most even.

小开

$ split -l 100 input_file output_file

where -l is the number of lines in each files. This will create:

output_fileaa
output_fileab
output_fileac
output_filead
....

小开

CS Pei's answer won't produce .txt files as the OP wants. Use:

split -b=1M -d  file.txt file --additional-suffix=.txt

小开

Regardless to what was said in previous answers, on my Ubuntu 16.04 (Xenial Xerus) I had to do:

split -b 10M -d  system.log system_split.log

Please note the space between -b and the value.

小开

Try something like this:

awk -vc=1 'NR%1000000==0{++c}{print $0 > c".txt"}' Datafile.txt


for filename in *.txt; do mv "$filename" "Prefix_$filename"; done;

小开

I agree with @CS Pei, however this didn't work for me:

split -b=1M -d file.txt file

...as the = after -b threw it off. Instead, I simply deleted it and left no space between it and the variable, and used lowercase "m":

split -b1m -d file.txt file

And to append ".txt", we use what @schoon said:

split -b=1m -d file.txt file --additional-suffix=.txt

I had a 188.5MB txt file and I used this command [but with -b5m for 5.2MB files], and it returned 35 split files all of which were txt files and 5.2MB except the last which was 5.0MB. Now, since I wanted my lines to stay whole, I wanted to split the main file every 1 million lines, but the split command didn't allow me to even do -100000 let alone "-1000000, so large numbers of lines to split will not work.

小开

On my Linux system (Red Hat Enterprise 6.9), the split command does not have the command-line options for either -n or --additional-suffix.

Instead, I've used this:

split -d -l NUM_LINES really_big_file.txt split_files.txt.

where -d is to add a numeric suffix to the end of the split_files.txt. and -l specifies the number of lines per file.

For example, suppose I have a really big file like this:

$ ls -laF
total 1391952
drwxr-xr-x 2 user.name group         40 Sep 14 15:43 ./
drwxr-xr-x 3 user.name group       4096 Sep 14 15:39 ../
-rw-r--r-- 1 user.name group 1425352817 Sep 14 14:01 really_big_file.txt

This file has 100,000 lines, and I want to split it into files with at most 30,000 lines. This command will run the split and append an integer at the end of the output file pattern split_files.txt..

$ split -d -l 30000 really_big_file.txt split_files.txt.

The resulting files are split correctly with at most 30,000 lines per file.

$ ls -laF
total 2783904
drwxr-xr-x 2 user.name group        156 Sep 14 15:43 ./
drwxr-xr-x 3 user.name group       4096 Sep 14 15:39 ../
-rw-r--r-- 1 user.name group 1425352817 Sep 14 14:01 really_big_file.txt
-rw-r--r-- 1 user.name group  428604626 Sep 14 15:43 split_files.txt.00
-rw-r--r-- 1 user.name group  427152423 Sep 14 15:43 split_files.txt.01
-rw-r--r-- 1 user.name group  427141443 Sep 14 15:43 split_files.txt.02
-rw-r--r-- 1 user.name group  142454325 Sep 14 15:43 split_files.txt.03




$ wc -l *.txt*
100000 really_big_file.txt
30000 split_files.txt.00
30000 split_files.txt.01
30000 split_files.txt.02
10000 split_files.txt.03
200000 total

小开

If each part has the same number of lines, for example 22, here is my solution:

split --numeric-suffixes=2 --additional-suffix=.txt -l 22 file.txt file

And you obtain file2.txt with the first 22 lines, file3.txt the 22 next line, etc.

Thank @hamruta-takawale, @dror-s and @stackoverflowuser2010

小开

My search of how to do this led me here, so I'm posting this here for others too:

To get all of the contents of the file, split is the right answer! But, for those looking to just extract a piece of a file, as a sample of the file, use head or tail:

# extract just the **first** 100000 lines of /var/log/syslog into
# ~/syslog_sample.txt
head -n 100000 /var/log/syslog > ~/syslog_sample.txt


# extract just the **last** 100000 lines of /var/log/syslog into
# ~/syslog_sample.txt
tail -n 100000 /var/log/syslog > ~/syslog_sample.txt