Differences between revisions 28 and 29
Revision 28 as of 2009-12-21 10:26:07
Size: 7085
Editor: s5591aacb
Comment:
Revision 29 as of 2009-12-21 13:48:10
Size: 7657
Editor: GreyCat
Comment: why IFS=$'\n' backfires
Deletions are marked like this. Additions are marked like this.
Line 134: Line 134:
=== Using for instead of while read ===
Line 155: Line 157:
}}}
Line 156: Line 159:
However, as discussed in [[BashFAQ/005|FAQ #5]], the use of `IFS=$'\n'` (or any other "whitespace" in `IFS`) causes the shell to consolidate all consecutive instances of the whitespace delimiter into one. In other words, it skips over blank lines. Thus,

{{{
    ~$ cat foo
    line one

    line three
    ~$ IFS=$'\n'; set -f; for line in $(<foo); do echo "$line"; done; unset IFS; set +f
    line one
    line three
    ~$
Line 157: Line 171:

If preservation of blank lines is important, just go back to using `while read`.

How can I read a file (data stream, variable) line-by-line?

Use a while loop and the read command:

    while read -r line
    do
        echo "$line"
    done < "$file"

The -r option to read prevents backslash interpretation (usually used as a backslash newline pair, to continue over multiple lines). Without this option, any backslashes in the input will be discarded. You should always use the -r option with read.

BASH can also iterate over the lines in a variable using a "here string":

    while read -r line; do
        echo "$line"
    done <<< "$var"

If your data source is the script's standard input, then you don't need any redirection at all.

If you want to operate on individual fields within each line, you may supply additional variables to read:

    # Input file has 3 columns separated by white space.
    while read -r first_name last_name phone; do
      ...
    done < "$file"

If the field delimiters are not whitespace, you can set IFS (input field separator):

    while IFS=: read -r user pass uid gid gecos home shell; do
      ...
    done < /etc/passwd

For TAB delimited files, use IFS=$'\t'.

Also, please note that you do not necessarily need to know how many fields each line of input contains. If you supply more variables than there are fields, the extra variables will be empty. If you supply fewer, the last variable gets "all the rest" of the fields after the preceding ones are satisfied. For example,

    while read -r first_name last_name junk; do
      ...
    done <<< 'Bob Smith 123 Main Street Elk Grove Iowa 123-555-6789'
    # Inside the loop, first_name will contain "Bob", and
    # last_name will contain "Smith".  The variable "junk" holds
    # everything else.

The read command modifies each line read, e.g. by default it removes all leading whitespace characters (blanks, tab characters, ... -- basically any leading characters present in IFS). If that is not desired, the IFS variable has to be cleared:

    while IFS= read -r line
    do
        echo "$line"
    done < "$file"

Note that reading a file line by line this way is very slow for large files. Consider using e.g. AWK instead if you get performance problems.

One may also read from a command instead of a regular file:

    some command | while read -r line; do
       other commands
    done

This method is especially useful for processing the output of find with a block of commands:

    find . -print0 | while IFS= read -r -d $'\0' file; do
        mv "$file" "${file// /_}"
    done

This command reads one filename at a time from the file command and renames the file so that its spaces are replaced by underscores.

Note the usage of -print0 in the find command, which uses NUL bytes as filename delimiters, and -d $'\0' in the read command to instruct it to read all text into the file variable until it finds a NUL byte. By default, find and read delimit their input with newlines; however, since filenames can potentially contain newlines themselves, this default behaviour will split up those filenames with newlines and cause the loop body to fail. Additionally it is necessary to unset IFS, because otherwise read would strip trailing whitespace. See FAQ #20 for more details.

Using a pipe to send find's output into a while loop places the loop in a SubShell and may therefore cause problems later on if the commands inside the body of the loop attempt to set variables which need to be used outside the loop; in that case, see FAQ 24, or use ProcessSubstitution like:

    while read -r line; do
        other commands
    done < <(some command)

Sometimes it's useful to read a file into an array, one array element per line. You can do that with the following example:

    oIFS=$IFS IFS=$'\n' arr=($(< myfile)) IFS=$oIFS
    # Warning: breaks if lines contain "*" or similar

This temporarily changes the Input Field Separator to a newline, so that each line will be considered one field by read. Then it populates the array arr with the fields. Then it sets the IFS back to what it was before.

This same trick works on a stream of data as well as a file:

    oIFS=$IFS IFS=$'\n' arr=($(find . -type f)) IFS=$oIFS
    # Same warning as the previous example

Of course, this will blow up in your face if the filenames contain newlines; see FAQ 20 for hints on dealing with such filenames.

Both of these array-stuffing examples fail if the shell encounters a glob that matches files in the current directory as one of the input lines. Glob expansion can be disabled with set -f and then re-enabled afterward with set +f if needed. For more details on arrays, see FAQ 5. Moreover, since bash will treat sequences of IFS whitespace as a single character, if the input has empty lines (meaning that groups of two or more consecutive \n characters appear in the file), they will be lost. So, for example:

    $ cat myfile
    line1

    line2
    line3
    $ oIFS=$IFS IFS=$'\n' arr=($(< myfile)) IFS=$oIFS
    $ declare -p arr
    declare -a arr='([0]="line1" [1]="line2" [2]="line3")'

In the end, the safest way to read a file into an array is still to use a loop:

    i=0
    while IFS= read -r arr[i++]; do :;done < "$file"
    # or  <<< "$var"  to iterate over a variable

On the other hand, if the file lacks a trailing newline (such as /proc/$$/cmdline on Linux), the line will not be printed by a while read ... loop, as read returns a failure that aborts the while loop, thus failing to print the ultimate line. It does, however, store the contents of the partial line in the variable, so you can test whether there was such an unterminated line by checking whether the variable is non-empty at the end of the loop:

    # This does not work:
    echo -en 'line 1\ntruncated line 2' | while read -r line; do echo $line; done

    # This does not work either:
    echo -en 'line 1\ntruncated line 2' | while read -r line; do echo "$line"; done; [[ $line ]] && echo -n "$line"

    # This works:
    echo -en 'line 1\ntruncated line 2' | (while read -r line; do echo "$line"; done; [[ $line ]] && echo "$line")

For a discussion of why the second example above does not work as expected, see FAQ #24.

Using for instead of while read

Using for instead of while read ... is generally less efficient and more "featureful":

    $ cat afile
    ef gh
    *

    $ while read i ; do echo "$i" ; done < afile
    ef gh
    *

    $ for i in $(<afile) ; do echo "$i" ; done
    ef
    gh
    afile
    # the glob was expanded, and it looped per word.

    #workaround:
    $ IFS=$'\n' ;set -f ; for i in $(<afile) ; do echo "$i" ; done
    ef gh
    *

However, as discussed in FAQ #5, the use of IFS=$'\n' (or any other "whitespace" in IFS) causes the shell to consolidate all consecutive instances of the whitespace delimiter into one. In other words, it skips over blank lines. Thus,

    ~$ cat foo
    line one

    line three
    ~$ IFS=$'\n'; set -f; for line in $(<foo); do echo "$line"; done; unset IFS; set +f
    line one
    line three
    ~$ 

If preservation of blank lines is important, just go back to using while read.

BashFAQ/001 (last edited 2023-06-28 01:53:29 by larryv)