Extract specific words from text value - postgresql

Question

I have a query and a returned value that looks like this:
select properties->>'text' as snippet from table where id = 31;
snippet
-----------------------------------
There are many variations of passages of Lorem Ipsum available, but the majority have suffered alteration in some form, by injected humour, or randomised words which don't look even slightly believable.
(1 row)
This returns as I expect based on my query.
Is there a way that I can slice the returned text to only return words from position 5 to position 8 for example? Or alternatively, slice by character position which I will be able to use as a workaround?
I have tried using:
select properties->>'text'[0:13] as snippet from table where id = 31;
Which I hoped would return:
There are many
But it hasn't worked.
Is this possibly to slice a jsonb text field?

Answer 1

To "slice by character position", you can simply use the substr() function:
select substr(properties->>'text', 1, 15) as snippet
from the_table
where id = 31;
If you really want "words", you can split the text into an array using e.g. regexp_split_to_array. Once you have an array, you can use the slice syntax:
select (regexp_split_to_array(properties->>'text','\s+'))[5:8] as snippet
from the_table
where id = 31;
This returns an array, if you want it as a string, you can use array_to_string()
select array_to_string((regexp_split_to_array(properties->>'text','\s+'))[5:8],' ') as snippet
from the_table
where id = 31;
If you need that frequently, I would wrap it into a function:
create function extract_words(p_input text, p_start int, p_end int)
returns text
as
$$
select array_to_string((regexp_split_to_array(p_input,'\s+'))[p_start:p_end],' ');
$$
language sql
immutable;
Then the query is much easier to read:
select extract_words(properties->>'text', 5, 8) as snippet
from the_table
where id = 31;