NOTES ON AWK

Awk is a shell command. It can be executed on a file:

awk '<pattern1> { <action1> } ... <patternN> { <actionN> }' <optional-declarations> <file>

or with a redirection:

<unix-command> | awk '<pattern1> { <action1> } ... <patternN> { <actionN> }' <optional-declarations> 

Data can thus be either a text file, or the output of a shell command. These data are splitted in a series of records (by default, the lines of the file). The records are splitted in fields (by default, the words).

Actions are instructions to apply to data depending on the conditions given by <pattern>.

The default action is print, i.e. <pattern> ⇔ <pattern> { print }.

Conditions are boolean instructions selecting the data on which the actions will apply. They can also be BEGIN or END instructions.

By default, all the data are considered.

Optional variable declarations can be added at the end (e.g. change the value of FS, etc.).

$i

→ identify the i^th field of the record (1≤i≤N).

$0

→ identify the whole record.

RS (Record Separator)

→ delimiter to split the data into different records. By default, it is \n.

ORS (Output Record Separator)

→ similar to RS, but for outputs. By default, it is \n.

NR (Number of Records)

→ number of current lines, by default (1≤NR≤N). It is 0 for an empty file.

FS (Field Separator)

→ delimiter to split the records into fields, during the reading. By default,it is SPC.

OFS (Output Field Separator)

→ similar to FS, but for outputs. By default, it is SPC.

NF (Number of Fields)

→ number of fields in the current record (1≤NF≤N). It is 0 for a blank line.

0

→ false.

1 or non zero

→ true.

&&

→ and.

||

→ or.

!

→ not.

+<var>

→ true if <var> is a numerical variable.

print(<val1>, ... , <valN>)

→ write <val1> to <valN>, which can be the value of a variable, a field or a string between quotes.

printf("<format>",<var>)

→ where the format can be a combination of strings and instructions:

%d

→ integer, free format;

%<n>d

→ integer, on <n> columns;

%f

→ real, free format;

%<n>.<d>f

→ real with <d> decimals, on <n> columns;

%s

→ string, free format;

%<n>s

→ string, on <n> columns.

toupper(<var>)

→ convert <var> in upper case.

substr(<var>,<first-char>,<length>)

→ extract from the string <var> the characters between <first-char> and <first-char>+<length>.

split(<varIN>,<varOUT>,<sep>)

→ split the string <varIN> in a table <varOUT> according to the separator <sep>.

length(<var>)

→ number of characters in <var>.

Simple operations can be done on variables. In particular:

Initialization

→ <var>=<val> (by default, a variable is initialized to 0);

Incrementation

→ <var>++;

Arbitrary Incrementation

→ <var>+=<incr>.

Tables are associative. They work as a dict in Python: ARR[<var>] where <var> is not necessarily an index, but can be a field.

BEGIN { <action> }

→ perform <action> before the reading. This can used to initialize variables.

END { < action> }

→ perform <action> after writing.

Iterative schemes

→ { for (<arr1> in <arr2>) <action> }.

Decision schemes

→ { if (<cond1>) { <action1> } else if (<cond2>) { <action2> } else { <action3> }.