CSPLIT(1) User Commands CSPLIT(1)

csplit - split files based on context

csplit [-ks] [-f prefix] [-n number] file arg1... argn

The csplit utility reads the file named by the file operand, writes all or part of that file into other files as directed by the arg operands, and writes the sizes of the files.

The following options are supported:

-f prefix

Names the created files prefix00, prefix01, ..., prefixn. The default is xx00 ... xxn. If the prefix argument would create a file name exceeding 14 bytes, an error results. In that case, csplit exits with a diagnostic message and no files are created.


Leaves previously created files intact. By default, csplit removes created files if an error occurs.

-n number

Uses number decimal digits to form filenames for the file pieces. The default is 2.


Suppresses the output of file size messages.

The following operands are supported:


The path name of a text file to be split. If file is -, the standard input will be used.

The operands arg1 ... argn can be a combination of the following:


Create a file using the content of the lines from the current line up to, but not including, the line that results from the evaluation of the regular expression with offset, if any, applied. The regular expression rexp must follow the rules for basic regular expressions. Regular expressions can include the use of '\/' and '\%'. These forms must be properly quoted with single quotes, since "\" is special to the shell. The optional offset must be a positive or negative integer value representing a number of lines. The integer value must be preceded by + or . If the selection of lines from an offset expression of this type would create a file with zero lines, or one with greater than the number of lines left in the input file, the results are unspecified. After the section is created, the current line will be set to the line that results from the evaluation of the regular expression with any offset applied. The pattern match of rexp always is applied from the current line to the end of the file.


This operand is the same as /rexp/[offset], except that no file will be created for the selected section of the input file.


Create a file from the current line up to (but not including) the line number line_no. Lines in the file will be numbered starting at one. The current line becomes line_no.


Repeat operand. This operand can follow any of the operands described previously. If it follows a rexp type operand, that operand will be applied num more times. If it follows a line_no operand, the file will be split every line_no lines, num times, from that point.

An error will be reported if an operand does not reference a line between the current position and the end of the file.

See largefile(7) for the description of the behavior of csplit when encountering files greater than or equal to 2 Gbyte (2^31 bytes).

Example 1 Splitting and combining files

This example creates four files, cobol00...cobol03.

example% csplit -f cobol filename \
   '/procedure division/' /par5./ /par16./

After editing the split files, they can be recombined as follows:

example% cat cobol0[0−3] > filename

This example overwrites the original file.

Example 2 Splitting a file into equal parts

This example splits the file at every 100 lines, up to 10,000 lines. The -k option causes the created files to be retained if there are less than 10,000 lines; however, an error message would still be printed.

example% csplit -k filename 100 {99}

Example 3 Creating a file for separate C routines

If prog.c follows the normal C coding convention (the last line of a routine consists only of a } in the first character position), this example creates a file for each separate C routine (up to 21) in prog.c.

example% csplit -k prog.c '%main(%' '/^}/+1' {20}

See environ(7) for descriptions of the following environment variables that affect the execution of csplit: LANG, LC_ALL, LC_COLLATE, LC_CTYPE, LC_MESSAGES, and NLSPATH.

The following exit values are returned:


Successful completion.


An error occurred.

See attributes(7) for descriptions of the following attributes:

CSI Enabled
Interface Stability Standard

sed(1), split(1), attributes(7), environ(7), largefile(7), standards(7)

The diagnostic messages are self-explanatory, except for the following:

arg − out of range

The given argument did not reference a line between the current position and the end of the file.

December 4, 2003 OmniOS