Improved performance when decoding the entire set of rows with streamable JSON formats

[slvrtrn]

Summary

• Improved performance when decoding the entire set of rows with streamable JSON formats (such as JSONEachRow or JSONCompactEachRow) by calling the ResultSet.json() method. Depending on the dataset, it's between 10-15% (like cell_towers) and 40% (with large rows and very long strings) less execution time. NB: The actual streaming performance when consuming the ResultSet.stream() (which was fast) hasn't changed. Only the ResultSet.json() method used a suboptimal stream processing in some instances, and now ResultSet.json() just consumes the same stream transformer provided by the ResultSet.stream() method.

  Before:

  [1277 ms   ][JSONEachRow       ][SELECT * FROM large_strings ORDER BY id ASC LIMIT 100000                        ]
  [668 ms    ][JSONEachRow       ][SELECT * FROM cell_towers ORDER BY (radio, mcc, net, created) ASC LIMIT 500000  ]

  After:

  [737 ms    ][JSONEachRow       ][SELECT * FROM large_strings ORDER BY id ASC LIMIT 100000                        ]
  [578 ms    ][JSONEachRow       ][SELECT * FROM cell_towers ORDER BY (radio, mcc, net, created) ASC LIMIT 500000  ]

• Removed the outdated decode function.

• Updated exported types and doc entries for DataFormat.

• Fixed weird flakiness in expect-type assertions

[slvrtrn]

@mshustov

What change gave the performance boost?

The decode function did a few unneeded passes on a huge string. It is way less efficient than the stream transformer we already had in the ResultSet.

[slvrtrn]

This: https://github.com/ClickHouse/clickhouse-js/pull/253/files#diff-46e9eda9880b41c04095d911536817d628992924b2bdd3a96f66529dd3628a5bL101-L104

Unless JIT optimizes it out (it does not), it's at least one extra full pass, worst case two.