Hey Tony , Great matter ! I am planning to use real time data processing, can you please help me if timeseries is having any tangible benefits in comparsion to DynamoDB.? when we should prefer timeseries and when Dynamodb , we are trying to get a some use case done for smart factory , where real temperature , flow rate, pressure values comes, and later if we can visualize co relation
Can anybody explain the "CROSS JOIN UNNEST(...)" expression at minute 25.15 ? Is it because the output of "INTERPOLATE_FILL" is some kind of (time->value) tuple? Why do we need a cross join?
Hi Tony - do you think Timestream is suitable for storing stock market (symbolc, open, close, volume. etc. - BUT at daily frequency), Options chains for given symbol but again Daily feed ? Question is from following perspectives (1) the volume could be very high on daily basis (2) the timestamp is there but it is day timestamp (not like every second , the way examples are shows in Timestream videos) (3) if the data query pattern is not by day but could be across the days for a given symbol such that part of data may roll to magnetic store OR for filters on measures (meaning open_price > 100.01) Any pointers would be appreciated. Also if there is any forum/support where I can further discuss the use case?
I've to migrate few terabytes of timeseries data from mysql to QWS Timestream. But seems its is fetch query is very slow. Tested with few millions of records. Fetch query taking average 12 seconds time. My query is : SELECT (array_agg(measure_value::varchar ORDER BY time DESC))[1] as value, max(time) as reported_time FROM fluent.controller_data where config_id='75718' and time > '2020-07-28 09:23:30' and time
Does QuickSight integration with Timestream not support joining across tables.
Hey Tony , Great matter ! I am planning to use real time data processing, can you please help me if timeseries is having any tangible benefits in comparsion to DynamoDB.? when we should prefer timeseries and when Dynamodb , we are trying to get a some use case done for smart factory , where real temperature , flow rate, pressure values comes, and later if we can visualize co relation
Can anybody explain the "CROSS JOIN UNNEST(...)" expression at minute 25.15 ? Is it because the output of "INTERPOLATE_FILL" is some kind of (time->value) tuple? Why do we need a cross join?
Hi Tony - do you think Timestream is suitable for storing stock market (symbolc, open, close, volume. etc. - BUT at daily frequency), Options chains for given symbol but again Daily feed ? Question is from following perspectives (1) the volume could be very high on daily basis (2) the timestamp is there but it is day timestamp (not like every second , the way examples are shows in Timestream videos) (3) if the data query pattern is not by day but could be across the days for a given symbol such that part of data may roll to magnetic store OR for filters on measures (meaning open_price > 100.01)
Any pointers would be appreciated. Also if there is any forum/support where I can further discuss the use case?
I've to migrate few terabytes of timeseries data from mysql to QWS Timestream. But seems its is fetch query is very slow. Tested with few millions of records. Fetch query taking average 12 seconds time.
My query is :
SELECT (array_agg(measure_value::varchar ORDER BY time DESC))[1] as value, max(time) as reported_time FROM fluent.controller_data where config_id='75718' and time > '2020-07-28 09:23:30' and time
Always enjoy Tony’s presentations.
Sounds nice until you said append only...
pretty standard for time-series single-table DBs