Closed nicholasjhorton closed 8 years ago
There is also etl_transform
and etl_load
.
etl
is the constructor. @hadley likes constructors, so I am inclined to keep that.
The four main functions in the pipeline are:
etl_extract()
etl_transform()
etl_load()
etl_cleanup()
The two functions etl_create
and etl_update
are just wrappers to that pipeline. @cpsievert also favors junking the wrapping functions. I'm happy to consider re-naming either, but etl_populate
sounds more like a replacement for etl_load
than etl_create
.
If I'm understanding things correctly there are:
etl() etl_create() etl_update() etl_extract(), and etl_cleanup()
functions in the etl framework. Might it make sense to change etl() to etl_init() and etl_create() to etl_populate()? The others feel more self explanatory (but "etl()" by itself doesn't make sense).