...one of the most highly
regarded and expertly designed C++ library projects in the
world.
— Herb Sutter and Andrei
Alexandrescu, C++
Coding Standards
Reads a batch of rows.
template<
class SpanElementType,
class... StaticRow
,
class CompletionToken>
auto
async_read_some_rows(
static_execution_state< StaticRow... >& st,
span< SpanElementType > output,
diagnostics& diag,
CompletionToken&& token);
Reads a batch of rows of unspecified size into the storage given by
output
. At most output.size()
rows will be read. If the operation represented by st
has still rows to read, and output.size() > 0
, at least one row will be read.
Returns the number of read rows.
If there are no more rows, or st.should_read_rows() == false
, this function is a no-op and returns
zero.
The number of rows that will be read depends on the connection's buffer
size. The bigger the buffer, the greater the batch size (up to a maximum).
You can set the initial buffer size in connection
's
constructor, using buffer_params::initial_read_size
. The buffer
may be grown bigger by other read operations, if required.
Rows read by this function are owning objects, and don't hold any reference to the connection's internal buffers (contrary what happens with the dynamic interface's counterpart).
The type SpanElementType
must be the underlying row type for one of the types in the StaticRow
parameter pack (i.e., one
of the types in underlying_row_t<StaticRow>...
). The type must match the resultset
that is currently being processed by st
.
For instance, given static_execution_state<T1, T2>
, when reading rows for the second
resultset, SpanElementType
must exactly be underlying_row_t<T2>
. If this is not the case, a runtime
error will be issued.
This function can report schema mismatches.
The handler signature for this operation is void(boost::mysql::error_code, std::size_t)
.
The storage that output
references must be kept alive until the operation completes.