Redshift 在基于 CASE 语句的查询过程中将字符插入到字符串中

Question

I have a field with entries that can be 3 Letters 1,2 or 3 Numbers. EG. AAA1 AAA01 AAA001

I need to standardize the output in my query to always be 6 characters.

So I need a good case statement. Here is what I have tried.

CASE WHEN LENGTH(field1) = 4 THEN LEFT(field1,3)||00||RIGHT(field1,1) WHEN LENGTH(field1) = 5 THEN LEFT(field1,3)||0||RIGHT(field1,2) ELSE field1 as field1

This always resulted in the AAA1 coming out as AAA01 instead of AAA001

I tried this hoping for a different result:

CASE WHEN field1 ~ '[A-Z]{3}[0-9]{1}' THEN LEFT(field1,3)||00||RIGHT(field1,1) WHEN LENGTH(field1) = 5 THEN LEFT(field1,3)||0||RIGHT(field1,2) ELSE field1 END AS field1

that did nothing right resulting in the fields that were already correct to be incorrect.

So I messed with the REGEX and tried

CASE WHEN field1 ~ '^([A-Z]{3}[0-9]{1})$' THEN LEFT(field1,3)||00||RIGHT(field1,1) WHEN field1 ~ '^([A-Z]{3}[0-9]{2})$' THEN LEFT(field1,3)||0||RIGHT(field1,2) ELSE field1 END AS field1

which went right back to all the AAA1 resulting in AAA01 and the rest being correct.

Now, I am doing a union and I am trying to dedupe the union as well with

WHERE NOT EXISTS (SELECT table2.field3 FROM table2 WHERE table1.field3 = table2.field3 AND table2.field10 IS NOT NULL)

I can't give the entire query or data as they are confidential data, but any suggestions would be great.

Answer 1

I don't think you need any case statement or regex for this. This is what I came up with:

select left(field,3)||lpad(substring(field,4),3,'0')

Answer 2

The answer is that apparently 00 is not something that it likes.

the following worked:

CASE WHEN field1 ~ '^([A-Z]{3}[0-9]{1})$' THEN LEFT(field1,3)||0||0||RIGHT(field1,1) WHEN field1 ~ '^([A-Z]{3}[0-9]{2})$' THEN LEFT(field1,3)||0||RIGHT(field1,2) ELSE field1 END AS field1