# GROUP BY Tags

# Query with GROUP BY Particular Tag Columns

SELECT entity, datetime, tags.file_system, tags.mount_point, AVG(value)
  FROM "df.disk_used"
WHERE datetime > NOW - 2*HOUR
  AND entity LIKE 'nurswgvml00%' AND tags.mount_point != '/'
  GROUP BY entity, tags.file_system, tags.mount_point, PERIOD(1 HOUR)

# Query with GROUP BY All Tag Columns

The query produces the same result as above assuming the list of series tags is the same, namely the following two series tags: mount_point and file_system.

Using columns tags and tags.* (in the SELECT expression) provides an option of grouping by tags without knowing available series tags ahead of time.

SELECT entity, datetime, tags.*, AVG(value)
  FROM "df.disk_used"
WHERE datetime > NOW - 2*HOUR
  AND entity LIKE 'nurswgvml00%' AND tags.mount_point != '/'
  GROUP BY entity, tags, PERIOD(1 HOUR)

# Results

| entity       | datetime                 | tags.file_system           | tags.mount_point | AVG(value)   |
|--------------|--------------------------|----------------------------|------------------|--------------|
| nurswgvml006 | 2016-08-16T10:00:00.000Z | //u113452.nurstr003/backup | /mnt/u113452     | 1698750587.0 |
| nurswgvml006 | 2016-08-16T11:00:00.000Z | //u113452.nurstr003/backup | /mnt/u113452     | 1698750587.0 |
| nurswgvml007 | 2016-08-16T10:00:00.000Z | //u113452.nurstr003/backup | /mnt/u113452     | 1698750587.0 |
| nurswgvml007 | 2016-08-16T11:00:00.000Z | //u113452.nurstr003/backup | /mnt/u113452     | 1698750587.0 |
| nurswgvml009 | 2016-08-16T10:00:00.000Z | /dev/sdb1                  | /opt             | 31156676.3   |
| nurswgvml009 | 2016-08-16T11:00:00.000Z | /dev/sdb1                  | /opt             | 31156757.7   |
| nurswgvml006 | 2016-08-16T10:00:00.000Z | /dev/sdc1                  | /media/datadrive | 50273598.6   |
| nurswgvml006 | 2016-08-16T11:00:00.000Z | /dev/sdc1                  | /media/datadrive | 50287589.8   |