sql how to combine three queries from two tables into one query

Question

i have the following two tables:

Table1

id  name
---------
A3  B2
A3  B400
A5  B100
A7  B200
A8  B6
A8  B2
A8  B3

and

Answer 1

Try with UNION DISTINCT like:

SELECT DISTINCT t1.id as ID,
        t2.company as Company,
        'FOUND' AS status
 FROM   table1 t1
        JOIN table2 t2
          ON t1.id = t2.id
group by ID

union distinct

SELECT DISTINCT t2.id as ID,
        t2.company as Company,
        'FOUND' AS status
 FROM   table1 t1
        JOIN table2 t2
          ON t1.name = t2.name
group by ID

union distinct

SELECT t1.name as ID,
        t1.name as Company,
        'NOT FOUND' AS status
 FROM   table1 t1
  WHERE  t1.name NOT IN (SELECT t2.name
                                FROM   table2 t2)
GROUP BY ID

Answer 2

I think you can use a query like the following:

SELECT DISTINCT IF(name2 IS NULL, name, ID) AS ID,
       IF(name2 IS NULL, name, Company) AS Company,
       IF(name2 IS NULL, 'NOT FOUND', 'FOUND') AS Status
FROM (       
   SELECT DISTINCT 
          CASE 
             WHEN t1.id = t2.id THEN t1.id 
             WHEN t1.name = t2.name THEN t2.id
             ELSE t1.id
          END AS ID, 
          CASE 
             WHEN t1.id = t2.id THEN t2.company
             WHEN t1.name = t2.name THEN t2.company
             ELSE t1.name
          END AS Company,
          t1.name,
          (SELECT Table2.name
           FROM Table2
           WHERE Table2.name = t1.name LIMIT 1) AS name2
   FROM Table1 AS t1
   LEFT JOIN Table2 AS t2 ON (t1.id = t2.id) OR (t1.name = t2.name)) AS t
ORDER BY ID;

The query use a single LEFT JOIN operation plus a correlated subquery.

Demo here

Answer 3

There's something unclear about your queries, because they shouldn't execute as you've shown them (due to select containing non-aggregates not in group by). But based on your explanation of what you're trying to do...

You could use outer joins and then use case logic and/or coalesce to determine which value to use in each case.

SELECT DISTINCT
       coalesce(t2_id.id, t2_name.id, t1.name) as ID
     , coalesce(t2_id.company, t2_name.company, t1.name) as Conpany
     , case when t2_id.id is not null or t2_name.name is not null
            then 'FOUND'
            else 'NOT FOUND'
       end status
  FROM           table1 t1
       LEFT JOIN table2 t2_id
              ON t1.id = t2_id.id
       LEFT JOIN table2 t2_name
              ON t1.name = t2_name.name

Note that I used DISTINCT to make sure the exact same row doesn't appear multiple times; but this may return multiple rows for an ID (with different Company values) depending on the data. I couldn't quite tell what was intended because the uses of DISTINCT and GROUP BY in the three queries as posted in the question didn't seem to add up for me.

Answer 4

These requirements are confusing enough it might be worth a re-assessment of your data model. I think the UNION solution is your best bet possibly modified to use UNION ALL for efficiency.

I did put together a mutex based hack that likely has as many subtle problems as any of the other queries on this page.

select
  coalesce(t2.id, t1.name) AS ID,
  coalesce(t2.company, t1.name) AS Company,
  if(isnull(t2.id), 'NOT FOUND', 'FOUND') as Status
from (select 0 as mutex union select 1) as m
left join table1 as t1 on 1 = 1
left join table2 as t2 on t1.name = t2.name or (t1.id = t2.id and mutex)
group by coalesce(t2.id, t1.name)

That said, please test these queries carefully and look over your data and results. There's a whole lot of room for error depending on your input data.