BOL

Awk is a programming language which allows easy manipulation of structured data and is mostly used for pattern scanning and processing. It searches one or more files to see if they contain lines that match with the specified patterns and then perform associated actions. The basic syntax is:

awk '/pattern1/ {Actions}
/pattern2/ {Actions}' file

The working of Awk is as follows
Awk reads the input files one line at a time.
For each line, it matches with given pattern in the given order, if matches performs the corresponding action.
If no pattern matches, no action will be performed.
In the above syntax, either search pattern or action are optional, But not both.
If the search pattern is not given, then Awk performs the given actions for each line of the input.
If the action is not given, print all that lines that matches with the given patterns which is the default action.
Empty braces with out any action does nothing. It wont perform default printing operation.
Each statement in Actions should be delimited by semicolon.
Say you have data.tsv with the following contents:

$ cat data/test.tsv
contig1 ACTGTCTGTCACTGTGTTGTGATGTTGTGTGTG
contig2 ACTTTATATATT
contig3 ACTTATATATATATA
contig4 ACTTATATATATATA
contig5 ACTTTATATATT
By default Awk prints every line from the file.

$ awk '{print;}' data/test.tsv
contig1 ACTGTCTGTCACTGTGTTGTGATGTTGTGTGTG
contig2 ACTTTATATATT
contig3 ACTTATATATATATA
contig4 ACTTATATATATATA
contig5 ACTTTATATATT
We print the line which matches the pattern contig3

$ awk '/contig3/' data/test.tsv
contig3 ACTTATATATATATA
Awk has number of builtin variables. For each record i.e line, it splits the record delimited by whitespace character by default and stores it in the $n variables. If the line has 5 words, it will be stored in $1, $2, $3, $4 and $5. $0 represents the whole line. NF is a builtin variable which represents the total number of fields in a record.

$ awk '{print $1","$2;}' data/test.tsv
contig1,ACTGTCTGTCACTGTGTTGTGATGTTGTGTGTG
contig2,ACTTTATATATT
contig3,ACTTATATATATATA
contig4,ACTTATATATATATA
contig5,ACTTTATATATT

$ awk '{print $1","$NF;}' data/test.tsv
contig1,ACTGTCTGTCACTGTGTTGTGATGTTGTGTGTG
contig2,ACTTTATATATT
contig3,ACTTATATATATATA
contig4,ACTTATATATATATA
contig5,ACTTTATATATT

Awk has two important patterns which are specified by the keyword called BEGIN and END. The syntax is as follows:

BEGIN { Actions before reading the file}
{Actions for everyline in the file}
END { Actions after reading the file }

For example,
$ awk 'BEGIN{print "Header,Sequence"}{print $1","$2;}END{print "-------"}' data/test.tsv
Header,Sequence
contig1,ACTGTCTGTCACTGTGTTGTGATGTTGTGTGTG
contig2,ACTTTATATATT
contig3,ACTTATATATATATA
contig4,ACTTATATATATATA
contig5,ACTTTATATATT
-------
We can also use the concept of a conditional operator in print statement of the form print CONDITION ? PRINT_IF_TRUE_TEXT : PRINT_IF_FALSE_TEXT. For example, in the code below, we identify sequences with lengths > 14:

$ awk '{print (length($2)>14) ? $0">14" : $0"<=14";}' data/test.tsv
contig1 ACTGTCTGTCACTGTGTTGTGATGTTGTGTGTG>14
contig2 ACTTTATATATT<=14
contig3 ACTTATATATATATA>14
contig4 ACTTATATATATATA>14
contig5 ACTTTATATATT<=14
We can also use 1 after the last block {} to print everything (1 is a shorthand notation for {print $0} which becomes {print} as without any argument print will print $0 by default), and within this block, we can change $0, for example to assign the first field to $0 for third line (NR==3), we can use:

$ awk 'NR==3{$0=$1}1' data/test.tsv
contig1 ACTGTCTGTCACTGTGTTGTGATGTTGTGTGTG
contig2 ACTTTATATATT
contig3
contig4 ACTTATATATATATA
contig5 ACTTTATATATT
You can have as many blocks as you want and they will be executed on each line in the order they appear, for example, if we want to print $1 three times (here we are using printf instead of print as the former doesn't put end-of-line character),

$ awk '{printf $1"\t"}{printf $1"\t"}{print $1}' data/test.tsv
contig1 contig1 contig1
contig2 contig2 contig2
contig3 contig3 contig3
contig4 contig4 contig4
contig5 contig5 contig5
Although, we can also skip executing later blocks for a given line by using next keyword:

$ awk '{printf $1"\t"}NR==3{print "";next}{print $1}' data/test.tsv
contig1 contig1
contig2 contig2
contig3
contig4 contig4
contig5 contig5

$ awk 'NR==3{print "";next}{printf $1"\t"}{print $1}' data/test.tsv
contig1 contig1
contig2 contig2

contig4 contig4
contig5 contig5
You can also use getline to load the contents of another file in addition to the one you are reading, for example, in the statement given below, the while loop will load each line from test.tsv into k until no more lines are to be read:

$ awk 'BEGIN{while((getline k <"data/test.tsv")>0) print "BEGIN:"k}{print}' data/test.tsv
BEGIN:contig1 ACTGTCTGTCACTGTGTTGTGATGTTGTGTGTG
BEGIN:contig2 ACTTTATATATT
BEGIN:contig3 ACTTATATATATATA
BEGIN:contig4 ACTTATATATATATA
BEGIN:contig5 ACTTTATATATT
contig1 ACTGTCTGTCACTGTGTTGTGATGTTGTGTGTG
contig2 ACTTTATATATT
contig3 ACTTATATATATATA
contig4 ACTTATATATATATA
contig5 ACTTTATATATT
You can also store data in the memory with the syntax VARIABLE_NAME[KEY]=VALUE which you can later use through for (INDEX in VARIABLE_NAME) command:

$ awk '{i[$1]=1}END{for (j in i) print j"<="i[j]}' data/test.tsv
contig1<=1
contig2<=1
contig3<=1
contig4<=1
contig5<=1