Optimize tag query by avoiding multi-query

[Arkatufus]

Changes

Use SQL built-in aggregate string to retrieve journal row and tags in 1 go

Latest dev Benchmarks

Tag Count	TagMode	Mean	Error	StdDev	Gen0	Gen1	Allocated
10	Csv	1,712.480 ms	22.7837 ms	21.3119 ms	-	-	390.94 KB
100	Csv	1,720.164 ms	21.3537 ms	19.9743 ms	-	-	1368.78 KB
1000	Csv	1,763.240 ms	30.8206 ms	28.8296 ms	1000.0000	-	10675.73 KB
10000	Csv	1,887.153 ms	35.8259 ms	35.1858 ms	12000.0000	3000.0000	103176.36 KB
10	TagTable	3.262 ms	0.0645 ms	0.1096 ms	42.9688	3.9063	355.72 KB
100	TagTable	4.935 ms	0.0969 ms	0.1420 ms	156.2500	31.2500	1307.93 KB
1000	TagTable	24.775 ms	0.4383 ms	0.3885 ms	1312.5000	656.2500	10923.02 KB
10000	TagTable	418.739 ms	5.3480 ms	5.0025 ms	12000.0000	3000.0000	106984.68 KB

This PR's Benchmarks

Tag Count	TagMode	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
10	Csv	1,760.393 ms	27.1970 ms	25.4401 ms	-	-	-	449.35 KB
100	Csv	1,766.355 ms	25.0182 ms	23.4021 ms	-	-	-	1368.33 KB
1000	Csv	1,755.960 ms	33.8171 ms	34.7276 ms	1000.0000	-	-	10733.67 KB
10000	Csv	1,905.026 ms	22.3564 ms	20.9122 ms	12000.0000	3000.0000	-	103175.1 KB
10	TagTable	2.336 ms	0.0389 ms	0.0344 ms	39.0625	3.9063	-	341.51 KB
100	TagTable	3.943 ms	0.0705 ms	0.0660 ms	148.4375	31.2500	-	1262.26 KB
1000	TagTable	18.597 ms	0.3570 ms	0.3506 ms	1281.2500	718.7500	-	10481.17 KB
10000	TagTable	184.446 ms	3.3447 ms	2.9650 ms	14000.0000	3666.6667	1666.6667	103154.93 KB

[Arkatufus]

Another optimization I tried is by passing the original journal query into the tag query like so:

var tagRows = await connection
    .GetTable<JournalTagRow>()
    .Where(r => r.OrderingId.In(rowQuery.Select(jr => jr.Ordering)))
    .Select(r => new TagRow
    {
        OrderingId = r.OrderingId,
        TagValue = r.TagValue
    })
    .ToListAsync();

This actually degrades the performance:

Tag Count	TagMode	Mean	Error	StdDev	Gen0	Gen1	Gen2	Allocated
10	Csv	1,752.93 ms	18.677 ms	15.596 ms	-	-	-	450.28 KB
100	Csv	1,722.95 ms	11.452 ms	8.941 ms	-	-	-	1368.69 KB
1000	Csv	1,726.12 ms	9.396 ms	8.789 ms	1000.0000	-	-	10734.09 KB
10000	Csv	1,895.89 ms	20.333 ms	19.020 ms	12000.0000	3000.0000	-	103175 KB
10	TagTable	20.60 ms	0.208 ms	0.194 ms	31.2500	-	-	379.39 KB
100	TagTable	23.35 ms	0.466 ms	0.518 ms	156.2500	31.2500	-	1320.66 KB
1000	TagTable	43.39 ms	0.850 ms	0.795 ms	1250.0000	666.6667	-	10836.95 KB
10000	TagTable	451.57 ms	7.244 ms	6.049 ms	13000.0000	4000.0000	1000.0000	105942.11 KB

[Arkatufus]

Use SQL built-in string aggregate function to retrieve all tags in one go

[Aaronontheweb]

@Arkatufus looks like this broke something across several Db implementations

[Arkatufus]

@Arkatufus looks like this broke something across several Db implementations

Should be fixed now

[Aaronontheweb]

LGTM