Steve McCanne
Steve McCanne
The unfinished vcache implementation currently loads vectors into memory and never evicts them. In this issue, we will extend the vcache API with ownership (so a query can release vectors...
The vcache currently loads the entire object into memory but should allow loading only the data needed by a query. This work involves implementing an interface to vcache to allow...
This work involves stubbing out the initial and simple Vector interface type for the vector engine and locating it in a package under the runtime. A rudimentary implementation will allow...
If you pipe the output of brimcap to zq and the zq query has an error in it, then the pipe fails silently because brimcap ignores EPIPE. It should exit...
To test the service defined in brimdata/zync#83, this issue is to create a performance test rig using large amounts of synthetic data. We will create a large number of data...
Syncing from kakfa and doing ETL should run as a continuous service so we don't need to poll and recompute progress state on each run. Step 1 is to get...
We should think through and implement how zync should do logging and change the various printf's to log calls. This will be used from orchestration and automation so these logs...
The yaml config should allow logic to refer to more than one output table and track the input cursors to multiple outputs.
The yaml config should allow logic to refer to more than two input tables to create complex joins between three or more inputs and track progress cursors accordingly.
In `etl/pipeline.go`, the commented out code cannot replace the code below for some reason... ``` //XXX This still doesn't work with the bug fix. See issue # //if vals[k].Type ==...