Sql-server – PaaS Server – Random row order while using aggregate function with group by

I initially thought you were on to something here. Working assumption was along the lines that perhaps the buffer pool wasn't immediately flushed as it requires "some work" to do so and why bother until the memory was required. But...

Your test is flawed.

What you're seeing in the buffer pool is the pages read as a result of re-attaching the database, not the remains of the previous instance of the database.

And we can see that the buffer pool was not totally blown away by the detach/attach. Seems like my buddy was wrong. Does anyone disagree or have a better argument?

Yes. You're interpreting physical reads 0 as meaning there were not any physical reads

Table 'DatabaseLog'. Scan count 1, logical reads 782, physical reads 0, read-ahead reads 768, lob logical reads 94, lob physical reads 4, lob read-ahead reads 24.

As described on Craig Freedman's blog the sequential read ahead mechanism tries to ensure that pages are in memory before they're requested by the query processor, which is why you see zero or a lower than expected physical read count reported.

When SQL Server performs a sequential scan of a large table, the storage engine initiates the read ahead mechanism to ensure that pages are in memory and ready to scan before they are needed by the query processor. The read ahead mechanism tries to stay 500 pages ahead of the scan.

None of the pages required to satisfy your query were in memory until read-ahead put them there.

As to why online/offline results in a different buffer pool profile warrants a little more idle investigation. @MarkSRasmussen might be able to help us out with that next time he visits.

Without knowing anything about the source data, perhaps this would do what you want?

USE Test;
GO
CREATE TABLE GENDER
(
    ORG INT NOT NULL
    , GENDER VARCHAR(1) NOT NULL
);

CREATE TABLE AGE
(
    ORG INT NOT NULL
    , AGE TINYINT
);

CREATE TABLE STATES
(
    ORG INT NOT NULL
    , STATENAME VARCHAR(255)
);

INSERT INTO Gender (ORG, GENDER) VALUES (1, 'M');
INSERT INTO Gender (ORG, GENDER) VALUES (1, 'F');
INSERT INTO Gender (ORG, GENDER) VALUES (2, 'M');
INSERT INTO Gender (ORG, GENDER) VALUES (2, 'F');
INSERT INTO Gender (ORG, GENDER) VALUES (3, 'M');
INSERT INTO Gender (ORG, GENDER) VALUES (3, 'F');

INSERT INTO AGE (ORG, AGE) VALUES (1,27);
INSERT INTO AGE (ORG, AGE) VALUES (1,28);
INSERT INTO AGE (ORG, AGE) VALUES (1,29);
INSERT INTO AGE (ORG, AGE) VALUES (1,30);
INSERT INTO AGE (ORG, AGE) VALUES (2,37);
INSERT INTO AGE (ORG, AGE) VALUES (2,38);
INSERT INTO AGE (ORG, AGE) VALUES (2,39);
INSERT INTO AGE (ORG, AGE) VALUES (2,40);
INSERT INTO AGE (ORG, AGE) VALUES (3, 2);

INSERT INTO STATES (ORG, STATENAME) VALUES (1,'FL');
INSERT INTO STATES (ORG, STATENAME) VALUES (1,'GA');
INSERT INTO STATES (ORG, STATENAME) VALUES (1,'MN');
INSERT INTO STATES (ORG, STATENAME) VALUES (1,'NM');
INSERT INTO STATES (ORG, STATENAME) VALUES (2,'FL');
INSERT INTO STATES (ORG, STATENAME) VALUES (2,'MN');
INSERT INTO STATES (ORG, STATENAME) VALUES (2,'NM');
INSERT INTO STATES (ORG, STATENAME) VALUES (3,'FL');
INSERT INTO STATES (ORG, STATENAME) VALUES (3,'GA');
INSERT INTO STATES (ORG, STATENAME) VALUES (3,'NM');

CREATE TABLE FACTS
(
    ORG INT NOT NULL
    , GENDER VARCHAR(1) NULL
    , AGE INT NULL
    , STATENAME VARCHAR(255) NULL
);

INSERT INTO FACTS (ORG, GENDER, AGE, STATENAME)
SELECT ORG, GENDER, NULL, NULL
FROM GENDER
GROUP BY ORG, GENDER
UNION ALL
SELECT ORG, NULL, AGE, NULL
FROM AGE
GROUP BY ORG, AGE
UNION ALL
SELECT ORG, NULL, NULL, STATENAME
FROM STATES;

SELECT *
FROM FACTS
ORDER BY ORG;

The results:

This will create a FACTS table that has all the data from several source tables. As @ypercube and @jon-seigel said, this really doesn't make much sense; perhaps we are missing something compelling about your setup.

If this is not what you were expecting, please provide the source tables, and any other pertinent details.