Hi.
Need some advice about how to stora my data.
It’s sizes of potatoes from a optical sorter, I get length, width and area of every potato that goes through the machine.
I read the data once a minute and get a count for every size that passed during the last minute
do i store it as counts with name for each size:
2 30mm
4 31mm
3 32mm
or do i store it as one number for eace time that size occurs:
30
30
31
31
31
31
32
32
32
Hi @pettaflex
I would suggest you store the data as:
|Timestamp | Length | Width | Area|
(what is the area of a potato…do you mean cross sectional area?)
Do the above so that Length, Width and Area are all recorded with the same timestamp.
Compute the count for every size (which can be any of the dimensions) using a Flux function.
The cameras look att the potato front and back so the cross section is all i get.
Other opical sorters for ex. carrots look att the product with 3 cameras and there you get a fairly good estimate of volume.
Thank you for the guidance
I don’t really understand what you are retrieving from the sorter.
At first you say length, width and area. But your example is a count of multiple “sizes”.
Anyway I would store it as follows formatted in Line protocol:
potatosorter,size=<size category>, count=<number of potatos> 1556813561098000000
And have tag combinations for every count. You really only store timeseries values in to fields, the tags are for grouping.
Alternatively you give every counter different fieldnames.
