如果数据每 30 秒出现一次，我如何按分钟或小时或天或周过滤数据，数据在某些时候不存在

Question

data is present is dynamodb and timestamp is integers and coming in the millisecond

I just wanted that how can I solve this, suppose the user ask for minutes wise data and average as an aggregation function for that minutes, and in every minute we are getting two data point but it may be possible there is no data for some point so for that case we have explicitly check and assign 0 for those points.

Answer 1

You cannot do functions such as aggregate on DynamoDB. You can get all of the items for the last timeframe so long as the timeframe is your sort-key.

Item Collection

PK	SK	Data
123	1674818685	mydata
123	1674818656	mydata
123	1674818632	mydata
123	1674818611	mydata
234	1674818600	mydata
234	1674818612	mydata
567	1674503932	mydata

Given the above schema, you want all of the data for item 123 for a 1 minute time frame, you can do so like below:

SELECT * FROM mytable WHERE PK = 123 AND SK BETWEEN time1 AND time2

---

Note that you can only get the results are only for a given partition key, if you the results for every item in the table you would need a Global Secondary Index.

All Items Using an Index

GSI_PK	SK	Data	PL
1	1674818685	mydata	123
1	1674818656	mydata	123
1	1674818632	mydata	123
1	1674818611	mydata	123
1	1674818600	mydata	123
1	1674818612	mydata	123
1	1674503932	mydata	123

SELECT * FROM mytable.myindex WHERE GSI_PK = 1 AND SK BETWEEN time1 AND time2

Please note, having a single item value for your GSI PK will limit your write requests to 1000 WCU, if your application requires more you will need to shard your index key .