CAQL Change Log
We will document changes to the CAQL language here.
Backwards compatibility is a major concern for us. Changes to existing behavior are rare and typically predicated on bug fixes or performance improvements. Breaking changes will be explicitly marked in this document.
2020-04-24
- Add
filter:*
package.
2020-02-03
- Add
integrate:while()
function. - Add broadcast support to
forecasting:slope()
andforecasting:regression()
. With this change, the functions will now forecast all input slots, not only the first one. - Add
forecasting:auto()
function. - Add
histogram:ratio_above()
andhistogram:ratio_below()
functions. - Fix a bug where
fill()
would not fill in data. - Improve performance of
window:merge()
. - Fix a bug where
find()
queries would fail during service restarts.
2019-12-02
- Add
histogram:subtract()
function. - Add
base
parameter tointegrate()
. - Add
#serial
directive. - Fixes to
NULL
literal.
2019-11-08
Allow multiple input streams to
delay()
.Performance improvements to
integrate()
anddelay()
functions.Deprecate
histogram:window
/histogram:rolling
in favour ofwindow:histogram
,window:rolling
.Revise the time aggregation functions
window:*
androlling:*
- Speedup processing by leveraging pre-aggregated data.
- Add support for multiple input streams.
- Align window boundaries consistently.
- Add
window:first()
function, that selects the first sample in each window. - Add
window:merge()
function, to aggregate histograms over time. - Add "skip" parameter to control the advancement of time windows.
- Add "period" parameter to control the granularity of input data.
- Add "align=start/end" parameter to control alignment of the output data.
- Add "offset" parameter to control window offset against UTC.
[Breaking Change] Change the default output of
window:*
in batch mode, to be synchronous with the data. Before this change the output ofwindow:*
function was delayed a full window size behind the input data. The old behaviour can be restored by passing align="end" as parameter.
2019-10-29
- Optimize a number of query patterns to leverage federated data processing:
find() | stats:{sum,mean}
find() | count()
find() | top()
find:histogram() | histogram:merge()
find:histogram() | histogram:sum() | stats:sum()
- Fix bugs with limiting and sorting outputs.
- Add default labels to
histogram:*
output. - Restrict sorting of results to the
find()
function, so that, e.g. top-k output is not sorted by label. - Add
tag:remove()
function. - Set default/max limits for CAQL find() queries to 1000/3000.
- Speed-up data fetching with the
metric()
,search:metric()
andmetriccluster()
functions. - Deprecate
search:metric()
andmetriccluster()
functions in favour offind()
. - Fix
count()
function, to not count NaN values.