The import benchmark will be much more useful if we can support multiple agents importing different data simultaneously. Here are some ideas for things agent number could control:
the set of columns being imported into
the set of rows being imported into
the frame being imported into
the index being imported into
Implementation would look something like:
Add a --agent-controls flag to cmd/import.go which is a string with four options (columns, rows, frame, index)
in bench/import.go modify the Init method to set up the benchmark's configuration based on the agent controls flag.
probably have a switch statement with four cases (and a default that returns error)
columns would change the base column id and the max column id to be offset by number of columns times agent number. Similar for rows.
frame and index would create a new frame or index by concatenating the frame or index name with the agent number
The import benchmark will be much more useful if we can support multiple agents importing different data simultaneously. Here are some ideas for things agent number could control:
Implementation would look something like:
--agent-controlsflag tocmd/import.gowhich is a string with four options (columns, rows, frame, index)bench/import.gomodify the Init method to set up the benchmark's configuration based on the agent controls flag.switchstatement with four cases (and a default that returns error)