com.phasmidsoftware.tableparser.json.table

circe codecs for Table[T] (they replace 1.x's TableJsonFormat, which used spray-json).

A Table is represented by a JSON object with a "rows" key (an array of the rows) and, if the table has a header, a "header" key (an array of the column names). For example:

{
 "rows" : [ { "playerA" : "Adam S", "playerB" : "Amy A" } ],
 "header" : [ "playerA", "playerB" ]
}

Use them with import com.phasmidsoftware.tableparser.json.table.TableCodecs.given. The rows need an Encoder (or Decoder): for a case class, derives Codec.AsObject gives both.

NOTE the header and the keys of the rows may differ. The header is the Table's own header: for a table parsed from CSV, that's the column names of the file (e.g. "director_name"). The keys of each row are whatever the row's Encoder writes: for a derived Codec, the field names (e.g. "directorName"). The Decoder reads the rows by their keys, and keeps the header just as it was, so a round trip loses nothing.