bug: Timestream feature differences #11554
Description
Is there an existing issue for this?
- I have searched the existing issues
Current Behavior
when running with a MULTI table, I haven seen several discrepancies that blocks us using localstack as the integration test tool.
In the test we wrote several records to the table that each has different columns, and tried to aggregate them by time. Then we noticed those differences after some try and failures:
- localstack doesn't support
AVGon aggregation. If there isAVGin the SQL selection, the result will always be empty. - for Timestream's real implementation, aggregation on all columns will return nullValue if the column doesn't exist with the record. However for localstack it will return "0.0" scalar values instead.
Expected Behavior
After writing those 2 records:
val aRecord = Record.builder().apply {
measureName("A")
time(timestamp)
timeUnit("MILLISECONDS")
dimensions(
listOf(
Dimension.builder().name("primary_key").value("key1").build(),
),
)
measureValueType(MeasureValueType.MULTI)
measureValues(
MeasureValue.builder().apply {
name("a_1")
value("1.0")
type(MeasureValueType.DOUBLE)
}.build(),
MeasureValue.builder().apply {
name("a_2")
value("0")
type(MeasureValueType.DOUBLE)
}.build(),
MeasureValue.builder().apply {
name("a_3")
value("0.5")
type(MeasureValueType.DOUBLE)
}.build(),
)
}.build()
val bRecord = Record.builder().apply {
measureName("B")
time(timestamp)
timeUnit("MILLISECONDS")
dimensions(
listOf(
Dimension.builder().name("primary_key").value("key2").build(),
),
)
measureValueType(MeasureValueType.MULTI)
measureValues(
MeasureValue.builder().apply {
name("b_1")
value("1.0")
type(MeasureValueType.DOUBLE)
}.build(),
MeasureValue.builder().apply {
name("b_2")
value("0")
type(MeasureValueType.DOUBLE)
}.build(),
MeasureValue.builder().apply {
name("b_3")
value("0.5")
type(MeasureValueType.DOUBLE)
}.build(),
)
}.build()I would expect the following SQL can be run:
SELECT primary_key, measure_name, bin(time, 15m) as time,
SUM(a_1) AS a_1, SUM(a_2) AS a_2, AVG(a_3) AS a_3,
SUM(b_1) AS b_1, SUM(b_2) AS b_2, AVG(b_3) AS b_3,
MIN(time) as start_time, MAX(time) as end_time, COUNT(*) AS count
FROM "db"."table"
WHERE time BETWEEN '2021-04-05 07:00:00' AND '2021-04-06 06:59:00'
AND primary_key IN ('key1', 'key2')
GROUP BY 1,2,3,4How are you starting LocalStack?
Custom (please describe below)
Steps To Reproduce
How are you starting localstack (e.g., bin/localstack command, arguments, or docker-compose.yml)
using testcontainer with kotlin
Client commands (e.g., AWS SDK code snippet, or sequence of "awslocal" commands)
awslocal s3 mb s3://mybucket
Environment
- OS:
- LocalStack:
LocalStack version:
LocalStack Docker image sha:
LocalStack build date:
LocalStack build git hash:Anything else?
No response