MySQL查询选择具有相同值的所有字段

Question

be I'm looking for help in writing a MySQL query.

Each row has a minimum of two rows with the same conversation_id. I want to select those that that have "delete" in the recipient_state field in all rows with the same conversation_id.

SELECT conversation_id
FROM xf_conversation_recipient
GROUP BY recipient_state HAVING (delete in all fields)

The following conversation_id would be selected

conversation_id recipient_state
1               delete
1               delete

The following conversation_id would NOT be selected

conversation_id recipient_state
1               delete
1               active

Answer 1

SELECT conversation_id, COUNT(DISTINCT recipient_state) AS nb, recipient_state
FROM xf_conversation_recipient
GROUP BY conversation_id
HAVING nb=1 AND recipient_state='delete'

The query groups by conversation_id, keeps only records having 1 distinct recipient_state and recipient_state is equal to 'delete'.

Answer 2

Best I can come up with is a self-left-join:

SELECT DISTINCT r1.conversation_id
FROM xf_conversation_recipient AS r1
LEFT JOIN xf_conversation_recipient AS r2 ON r2.conversation_id = r1.conversation_id AND r2.recipient_state != 'delete'
WHERE r1.recipient_state = 'delete' AND r2.conversation_id IS NULL

Basically, grab all rows where there are no matching rows that don't have "delete" as the state.

Answer 3

Given a table xf_conversation_recipient that looks like this:

+----------------+----------------+
|conversation_id |recipient_state |
+----------------+----------------+
|1               |delete          |
|1               |delete          |
|2               |active          |
|2               |delete          |
|3               |delete          |
|3               |delete          |
|4               |active          |
|4               |delete          |
|5               |active          |
|5               |active          |
|6               |delete          |
|6               |delete          |
+----------------+----------------+

The following query returns the IDs of all conversations matching your criteria

SELECT 
    conversation_id AS selectedId, 
    count(*) AS count    
FROM xf_conversation_recipient   
WHERE recipient_state = "delete" 
GROUP BY conversation_id 
HAVING count>1 

Returns:

+-----------+------+
|selectedId |count |
+-----------+------+
|1          |2     |
|3          |2     |
|6          |2     |
+-----------+------+

Note, depending on your application, you may want to stop here.

---

Nesting this query, we can extract only the selectedId column, and nesting it again we can use that as an IN condition on a query, like so:

SELECT * FROM xf_conversation_recipient 
WHERE xf_conversation_recipient.conversation_id IN (
    SELECT t1.selectedId FROM (
        SELECT 
            conversation_id AS selectedId, 
            count(*) AS count
        FROM xf_conversation_recipient
        WHERE recipient_state = "delete"
        GROUP BY conversation_id
        HAVING count>1
    ) t1  
)

Which returns:

+----------------+----------------+
|conversation_id |recipient_state |
+----------------+----------------+
|1               |delete          |
|1               |delete          |
|3               |delete          |
|3               |delete          |
|6               |delete          |
|6               |delete          |
+----------------+----------------+