是否可以使用mysql通过单个查询选择多个重叠的日期范围？

Question

I currently have a somewhat large SQL statement that UNIONs the same SELECT statement 3 times each time with a larger date range. IE Last 30 days, Last 90 Days and Last 180 Days. The SQL works just fine and I am even OK with the fact that the query takes quite a while to run, I am just wondering if there is anyway to do the same thing with just one SELECT statement.

Seeing as a couple of people asked to see my code:-

SELECT mdt_userid AS logging_officer,
    'Last 30 Days' AS timespan,
    ACD_Time,
    TRUNCATE (( 1 - ( (TotalAssisting+TotalBRB+TotalShift+TotalLunch+TotalBreak-IF(TotalLunch<(DaysLoggedIn*Lunch),TotalLunch,(DaysLoggedIn*Lunch))-IF((TotalBRB+TotalBreak)<(DaysLoggedIn*1800),(TotalBRB+TotalBreak),(DaysLoggedIn*1800))) / (TotalLoginTime-TotalEreq-TotalDefault) ) ) *100,2) AS Utilization_percent,
    TRUNCATE((GRS_total/calls)*100,1) AS Ratio,
    IF (TotalACW =0,0,TotalACW / calls) AS ACW,
    grs.percent AS percent,
    TRUNCATE (((grs_total - wo_ci) / grs_total) *100,2) AS percent_w_ci,
    TRUNCATE (calls / ( ((TotalLoginTime-TotalEreq-TotalDefault) /3600) / 8.75 ) ,2) AS call_per_hour,
    TRUNCATE (GRS_total / ( ((TotalLoginTime-TotalEreq-TotalDefault) /3600) / 8.75 ) ,2) AS grs_per_hour 
FROM (SELECT DISTINCT fullname,
            SUM( IF (ReasonCode =  '138',  `ReasonCodeDuration` , 0) ) AS TotalAssisting,
            SUM( IF (ReasonCode =  '146',  `ReasonCodeDuration` , 0) ) AS TotalBRB,
            SUM( IF (ReasonCode =  '143',  `ReasonCodeDuration` , 0) ) AS TotalShift,
            SUM( IF (ReasonCode =  '141',  `ReasonCodeDuration` , 0) ) AS TotalACW,
            SUM( IF (ReasonCode =  '231',  `ReasonCodeDuration` , 0) ) AS TotalEreq,
            SUM( IF (ReasonCode =  '0',  `ReasonCodeDuration` , 0) ) AS TotalDefault,
            COUNT(ReasonCode = '136') AS CountLunch,
            SUM( IF (ReasonCode =  '136',  `ReasonCodeDuration` , 0) ) AS TotalLunch,
            SUM( IF (ReasonCode =  '137',  `ReasonCodeDuration` , 0) ) AS TotalBreak,
            SUM(  `ReasonCodeDuration` ) AS NRduration
        FROM  `CiscoAgentNotready` 
        WHERE StartDate >= '".adjustdate($DateTo,0,0,-30)." 00:00:00' 
                AND StartDate <= '".$DateTo." 23:59:59'
        GROUP BY fullname)
        AS notready, 
    (SELECT DISTINCT agentname,
            SUM( loginduration ) AS TotalLoginTime, COUNT(Date(logondate)) as DaysLoggedIn
        FROM CiscoAgentLogintime
        WHERE logondate  >= '".adjustdate($DateTo,0,0,-30)." 00:00:00' 
                AND logondate  <= '".$DateTo." 23:59:59'
        GROUP BY agentname)
        AS logintime, 
    (SELECT DISTINCT full_name, 
            SUM( handled ) AS calls
        FROM  `CiscoAgentCalls` 
        WHERE DateCol   >= '".adjustdate($DateTo,0,0,-30)."'  
                AND DateCol   <= '".$DateTo."'
        GROUP BY full_name)
        AS calls, 
    (SELECT DISTINCT logging_officer, 
            SUM( first_call ) AS firstcall, 
            SUM(IF (Config_item =  'unknown', 1, 0) ) AS wo_ci,
            COUNT( * ) AS GRS_total, 
            TRUNCATE ((SUM( first_call ) / COUNT( * )) *100,2) AS percent
        FROM (SELECT logging_officer, 
                IF (logging_officer =  `Resolving_Officer` , 1, 0) AS first_call, 
                Config_item
            FROM  `callcentergrsdata` 
            WHERE log_Date >= '".adjustdate($DateTo,0,0,-30)." 00:00:00' 
                 AND log_Date <= '".$DateTo." 23:59:59') 
            AS a
        GROUP BY logging_officer) 
        AS grs,
    (SELECT DISTINCT agtName,
            SUM(TalkTime) as sumTalk,
            SUM(CallsHandled) as answered,
            SUM(TalkTime)/SUM(CallsHandled) as ACD_Time
        FROM (SELECT CONCAT( LastName,  ', ', FirstName ) AS agtName,
                TalkTime, 
                CallsHandled
            FROM `CiscoAgentAHT`
            WHERE DateCol   >= '".adjustdate($DateTo,0,0,-30)."'  
                AND DateCol   <= '".$DateTo."') as subAHT
        GROUP BY agtName) 
        AS AHT,
    techjtblnew
WHERE mdt_userid = '".$user."'
AND notready.fullname = logintime.agentname
AND calls.full_name = notready.fullname 
AND calls.full_name = AHT.agtName
AND techjtblnew.cisco = notready.fullname 
AND grs.logging_officer = techjtblnew.grs
UNION 
SELECT mdt_userid AS logging_officer,
    'Last 90 Days' AS timespan,
    ACD_Time,
    TRUNCATE (( 1 - ( (TotalAssisting+TotalBRB+TotalShift+TotalLunch+TotalBreak-IF(TotalLunch<(DaysLoggedIn*Lunch),TotalLunch,(DaysLoggedIn*Lunch))-IF((TotalBRB+TotalBreak)<(DaysLoggedIn*1800),(TotalBRB+TotalBreak),(DaysLoggedIn*1800))) / (TotalLoginTime-TotalEreq-TotalDefault) ) ) *100,2) AS Utilization_percent,
    TRUNCATE((GRS_total/calls)*100,1) AS Ratio,
    IF (TotalACW =0,0,TotalACW / calls) AS ACW,
    grs.percent AS percent,
    TRUNCATE (((grs_total - wo_ci) / grs_total) *100,2) AS percent_w_ci,
    TRUNCATE (calls / ( ((TotalLoginTime-TotalEreq-TotalDefault) /3600) / 8.75 ) ,2) AS call_per_hour,
    TRUNCATE (GRS_total / ( ((TotalLoginTime-TotalEreq-TotalDefault) /3600) / 8.75 ) ,2) AS grs_per_hour 
FROM (SELECT DISTINCT fullname,
            SUM( IF (ReasonCode =  '138',  `ReasonCodeDuration` , 0) ) AS TotalAssisting,
            SUM( IF (ReasonCode =  '146',  `ReasonCodeDuration` , 0) ) AS TotalBRB,
            SUM( IF (ReasonCode =  '143',  `ReasonCodeDuration` , 0) ) AS TotalShift,
            SUM( IF (ReasonCode =  '141',  `ReasonCodeDuration` , 0) ) AS TotalACW,
            SUM( IF (ReasonCode =  '231',  `ReasonCodeDuration` , 0) ) AS TotalEreq,
            SUM( IF (ReasonCode =  '0',  `ReasonCodeDuration` , 0) ) AS TotalDefault,
            COUNT(ReasonCode = '136') AS CountLunch,
            SUM( IF (ReasonCode =  '136',  `ReasonCodeDuration` , 0) ) AS TotalLunch,
            SUM( IF (ReasonCode =  '137',  `ReasonCodeDuration` , 0) ) AS TotalBreak,
            SUM(  `ReasonCodeDuration` ) AS NRduration
        FROM  `CiscoAgentNotready` 
        WHERE StartDate >= '".adjustdate($DateTo,0,0,-90)." 00:00:00' 
                AND StartDate <= '".$DateTo." 23:59:59'
        GROUP BY fullname)
        AS notready, 
    (SELECT DISTINCT agentname,
            SUM( loginduration ) AS TotalLoginTime, COUNT(Date(logondate)) as DaysLoggedIn
        FROM CiscoAgentLogintime
        WHERE logondate  >= '".adjustdate($DateTo,0,0,-90)." 00:00:00' 
                AND logondate  <= '".$DateTo." 23:59:59'
        GROUP BY agentname)
        AS logintime, 
    (SELECT DISTINCT full_name, 
            SUM( handled ) AS calls
        FROM  `CiscoAgentCalls` 
        WHERE DateCol   >= '".adjustdate($DateTo,0,0,-90)."'  
                AND DateCol   <= '".$DateTo."'
        GROUP BY full_name)
        AS calls, 
    (SELECT DISTINCT logging_officer, 
            SUM( first_call ) AS firstcall, 
            SUM(IF (Config_item =  'unknown', 1, 0) ) AS wo_ci,
            COUNT( * ) AS GRS_total, 
            TRUNCATE ((SUM( first_call ) / COUNT( * )) *100,2) AS percent
        FROM (SELECT logging_officer, 
                IF (logging_officer =  `Resolving_Officer` , 1, 0) AS first_call, 
                Config_item
            FROM  `callcentergrsdata` 
            WHERE log_Date >= '".adjustdate($DateTo,0,0,-90)." 00:00:00' 
                 AND log_Date <= '".$DateTo." 23:59:59') 
            AS a
        GROUP BY logging_officer) 
        AS grs,
    (SELECT DISTINCT agtName,
            SUM(TalkTime) as sumTalk,
            SUM(CallsHandled) as answered,
            SUM(TalkTime)/SUM(CallsHandled) as ACD_Time
        FROM (SELECT CONCAT( LastName,  ', ', FirstName ) AS agtName,
                TalkTime, 
                CallsHandled
            FROM `CiscoAgentAHT`
            WHERE DateCol   >= '".adjustdate($DateTo,0,0,-90)."'  
                AND DateCol   <= '".$DateTo."') as subAHT
        GROUP BY agtName) 
        AS AHT,
    techjtblnew
WHERE mdt_userid = '".$user."'
AND notready.fullname = logintime.agentname
AND calls.full_name = notready.fullname 
AND calls.full_name = AHT.agtName
AND techjtblnew.cisco = notready.fullname 
AND grs.logging_officer = techjtblnew.grs
UNION 
SELECT mdt_userid AS logging_officer,
    'Last 120 Days' AS timespan,
    ACD_Time,
    TRUNCATE (( 1 - ( (TotalAssisting+TotalBRB+TotalShift+TotalLunch+TotalBreak-IF(TotalLunch<(DaysLoggedIn*Lunch),TotalLunch,(DaysLoggedIn*Lunch))-IF((TotalBRB+TotalBreak)<(DaysLoggedIn*1800),(TotalBRB+TotalBreak),(DaysLoggedIn*1800))) / (TotalLoginTime-TotalEreq-TotalDefault) ) ) *100,2) AS Utilization_percent,
    TRUNCATE((GRS_total/calls)*100,1) AS Ratio,
    IF (TotalACW =0,0,TotalACW / calls) AS ACW,
    grs.percent AS percent,
    TRUNCATE (((grs_total - wo_ci) / grs_total) *100,2) AS percent_w_ci,
    TRUNCATE (calls / ( ((TotalLoginTime-TotalEreq-TotalDefault) /3600) / 8.75 ) ,2) AS call_per_hour,
    TRUNCATE (GRS_total / ( ((TotalLoginTime-TotalEreq-TotalDefault) /3600) / 8.75 ) ,2) AS grs_per_hour 
FROM (SELECT DISTINCT fullname,
            SUM( IF (ReasonCode =  '138',  `ReasonCodeDuration` , 0) ) AS TotalAssisting,
            SUM( IF (ReasonCode =  '146',  `ReasonCodeDuration` , 0) ) AS TotalBRB,
            SUM( IF (ReasonCode =  '143',  `ReasonCodeDuration` , 0) ) AS TotalShift,
            SUM( IF (ReasonCode =  '141',  `ReasonCodeDuration` , 0) ) AS TotalACW,
            SUM( IF (ReasonCode =  '231',  `ReasonCodeDuration` , 0) ) AS TotalEreq,
            SUM( IF (ReasonCode =  '0',  `ReasonCodeDuration` , 0) ) AS TotalDefault,
            COUNT(ReasonCode = '136') AS CountLunch,
            SUM( IF (ReasonCode =  '136',  `ReasonCodeDuration` , 0) ) AS TotalLunch,
            SUM( IF (ReasonCode =  '137',  `ReasonCodeDuration` , 0) ) AS TotalBreak,
            SUM(  `ReasonCodeDuration` ) AS NRduration
        FROM  `CiscoAgentNotready` 
        WHERE StartDate >= '".adjustdate($DateTo,0,0,-120)." 00:00:00' 
                AND StartDate <= '".$DateTo." 23:59:59'
        GROUP BY fullname)
        AS notready, 
    (SELECT DISTINCT agentname,
            SUM( loginduration ) AS TotalLoginTime, COUNT(Date(logondate)) as DaysLoggedIn
        FROM CiscoAgentLogintime
        WHERE logondate  >= '".adjustdate($DateTo,0,0,-120)." 00:00:00' 
                AND logondate  <= '".$DateTo." 23:59:59'
        GROUP BY agentname)
        AS logintime, 
    (SELECT DISTINCT full_name, 
            SUM( handled ) AS calls
        FROM  `CiscoAgentCalls` 
        WHERE DateCol   >= '".adjustdate($DateTo,0,0,-120)."'  
                AND DateCol   <= '".$DateTo."'
        GROUP BY full_name)
        AS calls, 
    (SELECT DISTINCT logging_officer, 
            SUM( first_call ) AS firstcall, 
            SUM(IF (Config_item =  'unknown', 1, 0) ) AS wo_ci,
            COUNT( * ) AS GRS_total, 
            TRUNCATE ((SUM( first_call ) / COUNT( * )) *100,2) AS percent
        FROM (SELECT logging_officer, 
                IF (logging_officer =  `Resolving_Officer` , 1, 0) AS first_call, 
                Config_item
            FROM  `callcentergrsdata` 
            WHERE log_Date >= '".adjustdate($DateTo,0,0,-120)." 00:00:00' 
                 AND log_Date <= '".$DateTo." 23:59:59') 
            AS a
        GROUP BY logging_officer) 
        AS grs,
    (SELECT DISTINCT agtName,
            SUM(TalkTime) as sumTalk,
            SUM(CallsHandled) as answered,
            SUM(TalkTime)/SUM(CallsHandled) as ACD_Time
        FROM (SELECT CONCAT( LastName,  ', ', FirstName ) AS agtName,
                TalkTime, 
                CallsHandled
            FROM `CiscoAgentAHT`
            WHERE DateCol   >= '".adjustdate($DateTo,0,0,-120)."'  
                AND DateCol   <= '".$DateTo."') as subAHT
        GROUP BY agtName) 
        AS AHT,
    techjtblnew
WHERE mdt_userid = '".$user."'
AND notready.fullname = logintime.agentname
AND calls.full_name = notready.fullname 
AND calls.full_name = AHT.agtName
AND techjtblnew.cisco = notready.fullname 
AND grs.logging_officer = techjtblnew.grs

Answer 1

Yes. Just query for the last 180 days and parse it in the software layer.

If you want an indicator of which one each row is, you can use CASE or IF statements:

http://dev.mysql.com/doc/refman/5.0/en/if-statement.html

IF search_condition THEN statement_list
    [ELSEIF search_condition THEN statement_list] ...
    [ELSE statement_list]
END IF

http://dev.mysql.com/doc/refman/5.0/en/case-statement.html

CASE case_value
    WHEN when_value THEN statement_list
    [WHEN when_value THEN statement_list] ...
    [ELSE statement_list]
END CASE

UPDATE

Ok, now that I know what you are going for I can provide a better solution that I believe will fit your needs.

I don't think that there is a way out of doing UNION BUT you if you do the UNION ahead of time adding a column for the various time spans you can just query once on the results.

In this example, I create a column called span during the UNION . The t1 is the resulting "table". Since span is different for each date range, the rows will show up multiple times.

Example Table:

mysql> select * from test;
+----+------------+
| id | when       |
+----+------------+
|  1 | 2012-02-01 |
|  2 | 2012-01-01 |
|  3 | 2011-09-01 |
+----+------------+

Query With Unions:

mysql> 
    ->    SELECT *, 1 as 'span' FROM `test` WHERE `when` > '2012-01-15'
    -> UNION
    ->    SELECT *, 2 as 'span' FROM `test` WHERE `when` > '2011-12-15'
    -> UNION
    ->    SELECT *, 3 as 'span' FROM `test` WHERE `when` > '2011-08-15';

+----+------------+------+
| id | when       | span |
+----+------------+------+
|  1 | 2012-02-01 |    1 |
|  2 | 2012-01-01 |    2 |
|  1 | 2012-02-01 |    2 |
|  3 | 2011-09-01 |    3 |
|  2 | 2012-01-01 |    3 |
|  1 | 2012-02-01 |    3 |
+----+------------+------+

Query Using That Union (custom SELECT and WHERE t1.id > 0 ):

mysql> 
    -> SELECT
    ->    `id`,
    ->    `when`,
    ->    if(`span` = 1, 'Last 30 Days',
    ->       if(`span` = 2, 'Last 60 Days',
    ->          'Last 180 Days')) as 'Timespan'
    -> FROM (
    ->       SELECT *, 1 as 'span' FROM `test` WHERE `when` > '2012-01-15'
    ->    UNION
    ->       SELECT *, 2 as 'span' FROM `test` WHERE `when` > '2011-12-15'
    ->    UNION
    ->       SELECT *, 3 as 'span' FROM `test` WHERE `when` > '2011-08-15'
    -> ) as t1
    -> WHERE t1.id > 0;

+----+------------+---------------+
| id | when       | Timespan      |
+----+------------+---------------+
|  1 | 2012-02-01 | Last 30 Days  |
|  2 | 2012-01-01 | Last 60 Days  |
|  1 | 2012-02-01 | Last 60 Days  |
|  3 | 2011-09-01 | Last 180 Days |
|  2 | 2012-01-01 | Last 180 Days |
|  1 | 2012-02-01 | Last 180 Days |
+----+------------+---------------+

You will do all of your query stuff off of t1 , not off the table itself. Adding the extra column in the creation of t1 allows each row to appear multiple times as the extra row will cause it to be unique.

Answer 2

Seems like you could get away with a complex grouping. Eg, here's a generic way to do it with records:

SELECT
  count(*),
  MIN(creation)
FROM record
WHERE creation > CURDATE() - INTERVAL 120 DAY
GROUP BY 
 IF(creation>CURDATE() - INTERVAL 30 DAY,'030',
   IF(creation>CURDATE() - INTERVAL 90 DAY,'090',
     IF(creation>CURDATE() - INTERVAL 120 DAY,'120',
       'OLDER'
     )
   )
 )

You would want to add the records as you go, as each range is exclusive.

Here's a sample output:

+----------+---------------+
| count(*) | MIN(creation) |
+----------+---------------+
|     1472 | 2012-01-08    |
|     2336 | 2011-11-09    |
|     1528 | 2011-10-10    |
|     5336 | 2011-10-10    |
+----------+---------------+
4 rows in set (0.13 sec)