High performance time series database
OVER TIME QQL issues: https://github.com/finos/TimeBase-CE/discussions/255
QQL Having Condition
HAVING can be used to filter results after GROUP BY or OVER TIME/COUNT. It can be useful for cases, if you need to filter a result but WHERE expression can affect this result.
Example:
WITH entries[this is TradeEntry] as 'entries'
SELECT first{}(entries[0].price) as 'open'
FROM "bitfinex"
OVER TIME(1m)
WHERE symbol == 'BTCUSD' AND notEmpty(entries) AND open is not null
In this case open expression is calculated during 'WHERE' expression (not after it!), so open will be calculated even if symbol is not 'BTCUSD' or entries are empty. This will lead to incorrect result.
WITH entries[this is TradeEntry] as 'entries'
SELECT first{}(entries[0].price) as 'open'
FROM "bitfinex"
OVER TIME(1m)
WHERE symbol == 'BTCUSD' AND notEmpty(entries)
HAVING open is not null
Another example:
SELECT count{}() as c
FROM "bars"
GROUP BY symbol
HAVING close > 100