Regex - Don't match nulls after words - regex

Question

I'm splitting a string on pipes (|) and the regex [^|]* is working fine when there are missing fields, but it's matching null characters after words:
GARBAGE|||GA|30604
yields matches
GARBAGE, null, null, null, GA, null, 30604, null
I've also tried [^|]+ which yields matches
GARBAGE, GA, 30604
EDIT: What I want is GARBAGE, null, null, GA, 30604. And |||| would yield matches null, null, null, null, null.
I'm referencing matches by index, so how can I fix the regex so it matches field by field, including nulls if there is no other data in the field?

Answer 1

This is how split works. You should use a split type function.
There is always a bias split uses.
Your case is simple in that it splits on a single character, in normal cases a regex is not needed.
And in this case, using a regex, the bias cannot be achieved without lookahead/lookbehind.
# (?:^|(?<=\|))([^|]*)(?=\||$)
(?:
^ # BOS
| # or
(?<= \| ) # Pipe behind
)
( [^|]* ) # (1), Optional non-pipe chars
(?=
\| # Pipe ahead
| # or
$ # EOS
)

Answer 2

While not exactly what you want, perhaps you could turn it into rows and work with it that way:
select nvl(regexp_substr( str, '([^|]*)\|{0,1}', 1, level, 'i', 1 ), 'null') part
from ( select 'GARBAGE|||GA|30604' str from dual )
connect by level <= regexp_count( str, '\|' ) + 1;
Specify which row (field) by adding a where clause where level equals the row (field) you want:
select nvl(regexp_substr( str, '([^|]*)\|{0,1}', 1, level, 'i', 1 ), 'null') part
from ( select 'GARBAGE|||GA|30604' str from dual )
where level = 4
connect by level <= regexp_count( str, '\|' ) + 1;