如何計算在列中由管道在postgresql分隔的列中的鍵值對清除值

Question

我有一個varchar列，其中包含值，我想提取的關鍵財務值，並忽略像測試，abc的單詞。

「其他：爲Transaction1 |金融：保險|理財：保險|其他：地點|金融：測試|金融：ABC」

輸出：保險

請您能有所幫助。

Answer 1

with temp as (
    select unnest(string_to_array(t.str,'|')) as str 
    from (values('Other:transaction1|finance:insurance|finance:insurance|Other:Place|finance:testing|finance:abc')) as t(str) 
) 
select distinct substring(str from position(':' in str)+1 for length(str)) 
from temp 
where str ~'finance:' and --matching finance 
     str !~'testing' and -- ignore words testing 
     str !~'abc'  -- ignore words abc 

Answer 2

一種方法是使用regexp_matches，這樣的事情：

with the_table(id, col) as(
    select 1, 'Other:transaction1|finance:testing|finance:testing|finance:Place|finance:testing|finance:insurance'::text union all 
    select 2, 'finance:insurance'::text union all 
    select 3, 'ffinance:insurance|finance:testing'::text union all 
    select 4, 'Other:transaction1|finance:insurance|finance:insurance|Other:Place|finance:testing|finance:abc'::text 
) 

select distinct id, arr[1] as word from (
    select id, regexp_matches('|'||col, '\|finance:([^\|]+)', 'g') as arr 
    from the_table 
) t 
where 
arr[1] not in('testing', 'abc') -- there is words you don't want to see in result 
order by id 

這給除「黑名單」作爲不同的行每個單詞。我使用獨特的列id只是爲了顯示，哪些字對應哪一行。