.. _inline:

Inline Filtering
=================

.. _sed:

sed
----

.. index:: sed

:title:`Linux for Programmers and Users`, Section 4.12.2.

.. program:: sed
.. describe:: sed

    A programmable stream editor for filtering and transforming text.

SYNOPSIS

    :program:`sed` [-e *script*] [-f *scriptfile*] [FILE ...]

.. option:: -e script, --expression=script

    add the script with the commands to be executed. **Note:** When only one
    `sed` script is used, the `-e` option is not needed; however, when
    more than one `sed` script is needed, use the `-e` before each script.

.. option:: -f script-file, --file=script-file

    add the contents of script-file with the commands to be executed

.. option:: -i[SUFFIX], --in-place[=SUFFIX]

    edit files in place (makes backup if extension supplied)

The :command:`sed` command is often used as a one line script, but may also be
invoked with a file containing a multi-line script -- much like a programming
language.

Sed Script Commands
--------------------

The general format of :command:`sed` commands is::

    [address-spec]  command

Commands that operate on whole lines (insert or delete lines) require an
address (line) specification.  The 's' command, which substitutes text with a
line, may either operate on all of the lines processed or use an optional
address (line) specification.

    /*regex*/
        Apply the command to any line matching the regular expression.

    *n*
        Apply the command to line (number) *n* in the data.

    *start*, *stop* 
        Apply the command between lines *start* and *stop* in the data.

The text editing commands are as follows.  

    *address* **a**\ \\\ *text*
        Append line(s) of text after *address*

    *address* **c**\ \\\ *text*
        Replace line(s) of text

    *address* **d**
        Delete line(s)

    *address* **i**\ \\\ *text*
        Insert line(s) of text before *address*

    *address* **r** *filename*
        Append a file after the address.  Note: use :command:`cat` to insert
        a file before the first line.

    [*address*] **s**/*regex*/*string*/
        Replace the first instance of text matching the regular expression 
        on the applied lines with the *string*.
        
    [*address*] **s**/*regex*/*string*/**g**
        Replace all instances of text matching the regular expression 
        on the applied lines with the *string*.


.. note::

    One common usage of sed is to generate edited string variable names in a
    shell script program.  For example::

        newname=$(echo $oldname | sed 's/regex/str/')
            
    See the text book for more examples.

.. _tr:

tr
---

:title:`Linux for Programmers and Users`, Section 4.12.3.

.. index:: tr
.. program:: tr
.. describe:: tr

    Replace characters from one sequence of characters with characters from
    another sequence.  The most common usage of :command:`tr` is to convert
    text data to either all capital letters or all lower case letters.

SYNOPSIS

    :command:`tr` [-cds] sequence1 sequence2

.. option:: -c, -C, --complement

    first complement SET1

.. option:: -d, --delete

    delete characters in SET1, do not translate

.. option:: -s, --squeeze-repeats

    replace each input sequence of  a  repeated  character  that  is
    listed in SET1 with a single occurrence of that character

Examples

Convert all letters in a string variable to lower case letters::

    input=$(echo $input | tr A-Z a-z)

:command:`tr` has some predefined characters sets that can be useful. See
the man page for the full list. One of them is ``[:space:]``, which we can
use to remove any white space or line feeds from a string that was read from
a file.  We also need some features of :ref:`echo` to make this work.  This
accomplishes the same thing as the `strip()` function in Python and
`chomp()` function in Perl. 

::

    line=$(echo -ne "${line}" | tr -d '[[:space:]]')