influxdb将部分tag转为field

参考

https://www.ancii.com/afkwpua8w/

https://blog.csdn.net/m0_37658349/article/details/120950353

 造数据：

insert instance,dc=us-west-2,server=ec2-dcas-address-db-01 cpu_usage_user=5.2,cpu_usage_system=1.0,cpu_usage_idle=90.7,cpu_usage_wait=3.1
insert instance,dc=us-west-2,server=ec2-dcas-address-db-01 memory_total=8009556,memory_used=1417488,memory_free=244076,memory_shared=118960,memory_buffered=0,memory_cached=6347992,memory_available=6082012
insert instance,dc=us-west-2,server=ec2-dcsindexsvc-01 cpu_usage_user=4.1,cpu_usage_system=2.0,cpu_usage_idle=97.7,cpu_usage_wait=2.3
insert instance,dc=us-west-2,server=ec2-dcsindexsvc-01 memory_total=32394732,memory_used=11330204,memory_free=5618464,memory_shared=229908,memory_buffered=0,memory_cached=15446064,memory_available=20433240
insert instance,dc=us-west-2,server=ec2-autodenalicontentpoi-01 cpu_usage_user=13.6,cpu_usage_system=0.8,cpu_usage_idle=85.6,cpu_usage_wait=0.6
insert instance,dc=us-west-2,server=ec2-autodenalicontentpoi-01 memory_total=64990084,memory_used=842448,memory_free=348420,memory_shared=17174872,memory_buffered=0,memory_cached=63799216,memory_available=46397536

1.  注意：

2. 插入的数据中measurement为instance，tag为dc和server，field为cpu***和memory***.
3. 上述数据中，虽然插入了6条数据，但是每两条数据的tag都是重复的，根据Series集合的定义，在插入的6条数据中，属于3个series。

查看当前数据库的Series信息

> show series;
key
---
instance,dc=us-west-2,server=ec2-autodenalicontentpoi-01
instance,dc=us-west-2,server=ec2-dcas-address-db-01
instance,dc=us-west-2,server=ec2-dcsindexsvc-01

 Series中的key为measurement,tag_key1=tag_value1,tag_key2=tag_value2..., value即为field_key1=field_value1,field_key2=field_value2...

现在的需求是： 

将 server索引转为filed，只留一个索引 dc

SELECT server, cpu_usage_user,cpu_usage_system,cpu_usage_idle,cpu_usage_wait,memory_total,memory_used,memory_free,memory_shared,memory_buffered,memory_cached,memory_available   INTO instancenew FROM instance4 group by dc

但是：  实际测试时，用这试，当数据量几十W时，发现select count() 数量有减少，不知什么原因。