Home / Questions / 409456
Accepted
Álvaro Rodríguez Pardo
Asked: 2021-11-25 06:36:25 +0800 CST 2021-11-25 06:36:25 +0800 CST

Using an awk script, how to get all the fields between two tokens based on any argument of a given field?

  • 772

Let's imagine that I have a file with the following fields, delimited by the token $$$$:

$$$$

> <DATABASE_ID>
HMDB0000016

> <DATABASE_NAME>
hmdb

> <FORMULA>
C21H30O3

$$$$

 <DATABASE_ID>
HMDB0000017

> <DATABASE_NAME>
hmdb

> <FORMULA>
C8H9NO4

$$$$

> <DATABASE_ID>
HMDB0000020

> <DATABASE_NAME>
hmdb

> <FORMULA>
C8H8O3

$$$$

In this case, I would like to be able to get all the fields that are between any two tokens using the value of the DATABASE_ID field via an awk script. For example, suppose I choose the value associated with this field HMDB0000020. How could I pass the value of this field to the awk script so that it prints on the screen all the fields that are between the tokens?

So far, I have managed to write a command from the command line that prints on the screen what is between the two tokens. I have achieved this command thanks to the post https://es.stackoverflow.com/a/407481/83.

Continuing with the example:

awk -v RS='\n\\$\\$\\$\\$\n' '/HMDB0000020/' fichero

> <DATABASE_ID>
HMDB0000020
> <DATABASE_NAME>
hmdb
> <FORMULA>
C8H8O3

However, I'm interested in being able to call a script using awk -f -v variable=HMDB0000020it and have the same output, but I don't know how I could rewrite the command I have.

Any ideas?

Thanks in advance.

linux

1 Answers

  1. Best Answer

    If you want to move the logic to an Awk file, you simply have to create a new file and, for better visibility, give it a .awk.

    For example, we can create "s.awk" with this content:

    BEGIN{
    RS="\n\\$\\$\\$\\$\n"
}
$0 ~ var

    And we execute it as follows:

    awk -f s.awk -v var="HMDB0000020" fichero

    That is, we execute Awk through the script s.awkand pass the variable to it varusing the syntax -v nombre_variable="valor variable", as we mentioned in How can I use shell variables in awk? . Finally, ficheroit is the content that you use to process.

    As for the content of the file itself, notice that I have moved the RS logic to the BEGIN block, which is the way we have to change this value without passing it when calling Awk "inline". That is, awk -v RS="bla"it is the same as awk 'BEGIN{RS="bla"}and allows us to put it inside the file.

    Then, the part of $0 ~ varconsists of validating that the given line has the same pattern that you pass to it through the variable "var".

    • 1