- BREAKING:
load-csv
argument order, meanings, and usage.- Only
csv-file
is required. - Database
cfg
instead ofconn
. - Ref- and tuple-related options separated out from the rest, with a map of schema-related options
:schema
,:ref-map
,:tuple-map
:composite-tuple-map
replacingcol-schema
.
- Only
- Full user specification of schema possible using alternative form for
:schema
to that for partial specification.
- Library name: from "datahike-csv-loader" to "tablehike"
- Support for chunking, for files too large to fit within memory
- BREAKING:
load-csv
argument order, semantics, and usage.- Database
cfg
has been supplanted byconn
. filename
andconn
are required.schema-opts
->schema-desc
, with different keys and semantics, as described in the docstring andREADME.md
.- Optional
parsers-desc
argument added for specification of column parsers / types.
- Database
- Type / Parser inference implemented within repo, instead of relying on
scicloj/tablecloth
.
tablehike.parse.parser
namespace added as replacement forscicloj/tablecloth
dependency (see last item in the list of changes for this release).tablehike.schema
namespace added for schema inference, in conjunction with parser / type inference implementation.
- Library name: from "tablehike" to "mesalog"
- BREAKING:
core
namespace renamed toapi
.