powered by
For a shard_table_buffer, this returns a lightweight in-memory handle (or a materialized data.frame/tibble, depending on materialize).
shard_table_buffer
materialize
table_finalize( target, materialize = c("never", "auto", "always"), max_bytes = 256 * 1024^2, ... )
A shard_table_handle, shard_row_groups, or materialized data.frame/tibble depending on target type and materialize.
shard_table_handle
shard_row_groups
target
A shard_table_buffer or shard_table_sink.
shard_table_sink
"never", "auto", or "always".
"never"
"auto"
"always"
For "auto", materialize only if estimated bytes <= max_bytes.
Reserved for future extensions.
For a shard_table_sink, this returns a row-group handle referencing the written partitions (or materializes them if requested).
# \donttest{ s <- schema(x = float64(), y = int32()) tb <- table_buffer(s, nrow = 5L) table_write(tb, idx_range(1, 5), data.frame(x = rnorm(5), y = 1:5)) handle <- table_finalize(tb) # }
Run the code above in your browser using DataLab