REGEX in Snowflake/SQL for Serialized Ruby Hash values - sql

Question

This is a tough one (for me at least). I need to obtain the words after "host_goal_to_have" and "host_goal_to_be_able" and there's not really a strong pattern to signify the end.
For example, the result for "host_goal_to_have" should be: "established a thriving community of actors supporting each other and turning passion into income!". And for "host_goal_to_be_able" being: "reach people and unlock the artist inside them"
Sample:
--- !ruby/hash:ActiveSupport::HashWithIndifferentAccess
signup_recaptcha:
hide_welcome_checklist: true
seen_spaces_migration_welcome_intro: true
host_goal_to_have: established a thriving community of actors supporting each other
and turning passion into income!
host_goals_updated_at: '2023-02-11T23:00:52.441Z'
host_goal_to_be_able: reach people and unlock the artist inside them
seen_event_form: true

Answer 1

So we can start with the same data:
with data as (
select * from values
($$--- !ruby/hash:ActiveSupport::HashWithIndifferentAccess
signup_recaptcha:
hide_welcome_checklist: true
seen_spaces_migration_welcome_intro: true
host_goal_to_have: established a thriving community of actors supporting each other
and turning passion into income!
host_goals_updated_at: '2023-02-11T23:00:52.441Z'
host_goal_to_be_able: reach people and unlock the artist inside them
seen_event_form: true$$)
)
and then we can split it into lines, find the matching "start of line tokens [a-z_]: and assume these are not embedded into a values section. Then separate token: value from value
select d.column1
,s.*
,RLIKE(s.value, '^[a-z_]+:.*$') as r1
,regexp_substr(s.value, '^[a-z_]+:(.*)$', 1,1,'c',1) as r2
,iff(r1, r2, s.value) as ss
from data as d,
table(split_to_table(d.column1, '\n')) as s
after this we can chain the values back together:
with data as (
select * from values
($$--- !ruby/hash:ActiveSupport::HashWithIndifferentAccess
signup_recaptcha:
hide_welcome_checklist: true
seen_spaces_migration_welcome_intro: true
host_goal_to_have: established a thriving community of actors supporting each other
and turning passion into income!
host_goals_updated_at: '2023-02-11T23:00:52.441Z'
host_goal_to_be_able: reach people and unlock the artist inside them
seen_event_form: true$$)
), pass_1 as (
select d.column1
,s.seq
,s.index
,regexp_substr(s.value, '^[a-z_]+:') as t
,nvl(t, lag(t)ignore nulls over(partition by s.seq order by s.index)) as token
,iff(RLIKE(s.value, '^[a-z_]+:.*$'), regexp_substr(s.value, '^[a-z_]+:(.*)$', 1,1,'c',1), s.value) as ss
from data as d,
table(split_to_table(d.column1, '\n')) as s
)
select
seq,
token,
listagg(ss) within group (order by index) as val
from pass_1
where token is not null
group by 1,2
;
which gives:
SEQ
TOKEN
VAL
1
seen_spaces_migration_welcome_intro:
true
1
host_goals_updated_at:
'2023-02-11T23:00:52.441Z'
1
signup_recaptcha:
1
host_goal_to_have:
established a thriving community of actors supporting each other and turning passion into income!
1
host_goal_to_be_able:
reach people and unlock the artist inside them
1
hide_welcome_checklist:
true
1
seen_event_form:
true
which can be filter via HAVING:
select
seq,
token,
trim(listagg(ss) within group (order by index)) as val
from pass_1
where token is not null
group by 1,2
having token in ('host_goal_to_have:', 'host_goal_to_be_able:')
SEQ
TOKEN
VAL
1
host_goal_to_have:
established a thriving community of actors supporting each other and turning passion into income!
1
host_goal_to_be_able:
reach people and unlock the artist inside them