result_data: include info about column data types

CassDataType can be a heavy object if the type is nested.
In this commit we include info about column data types in
result metadata (Vec<Arc<CassDataType>>).

This will be needed to implement cass_result_column_data_type function.

We also adjust the construction of `CassValue::value_type` field
and avoid allocations. We simply clone Arc from result metadata.

Author: muzarski

scylla-rust-wrapper/src/query_result.rs

@@ -1,6 +1,6 @@
 use crate::argconv::*;
 use crate::cass_error::CassError;
-use crate::cass_types::{cass_data_type_type, CassDataType, CassValueType};
+use crate::cass_types::{cass_data_type_type, get_column_type, CassDataType, CassValueType};
 use crate::inet::CassInet;
 use crate::metadata::{
     CassColumnMeta, CassKeyspaceMeta, CassMaterializedViewMeta, CassSchemaMeta, CassTableMeta,
@@ -22,9 +22,30 @@ pub struct CassResult {
 pub struct CassResultData {
     pub paging_state: Option<Bytes>,
     pub col_specs: Vec<ColumnSpec>,
+    pub col_data_types: Vec<Arc<CassDataType>>,
     pub tracing_id: Option<Uuid>,
 }
 
+impl CassResultData {
+    pub fn from_result_payload(
+        paging_state: Option<Bytes>,
+        col_specs: Vec<ColumnSpec>,
+        tracing_id: Option<Uuid>,
+    ) -> CassResultData {
+        let col_data_types = col_specs
+            .iter()
+            .map(|col_spec| Arc::new(get_column_type(&col_spec.typ)))
+            .collect();
+
+        CassResultData {
+            paging_state,
+            col_specs,
+            col_data_types,
+            tracing_id,
+        }
+    }
+}
+
 /// The lifetime of CassRow is bound to CassResult.
 /// It will be freed, when CassResult is freed.(see #[cass_result_free])
 pub struct CassRow {

scylla-rust-wrapper/src/session.rs

@@ -1,7 +1,7 @@
 use crate::argconv::*;
 use crate::batch::CassBatch;
 use crate::cass_error::*;
-use crate::cass_types::{get_column_type, CassDataType, UDTDataType};
+use crate::cass_types::{CassDataType, UDTDataType};
 use crate::cluster::build_session_builder;
 use crate::cluster::CassCluster;
 use crate::exec_profile::{CassExecProfile, ExecProfileName, PerStatementExecProfile};
@@ -204,11 +204,7 @@ pub unsafe extern "C" fn cass_session_execute_batch(
         match query_res {
             Ok(_result) => Ok(CassResultValue::QueryResult(Arc::new(CassResult {
                 rows: None,
-                metadata: Arc::new(CassResultData {
-                    paging_state: None,
-                    col_specs: vec![],
-                    tracing_id: None,
-                }),
+                metadata: Arc::new(CassResultData::from_result_payload(None, vec![], None)),
             }))),
             Err(err) => Ok(CassResultValue::QueryError(Arc::new(err))),
         }
@@ -289,11 +285,11 @@ pub unsafe extern "C" fn cass_session_execute(
 
         match query_res {
             Ok(result) => {
-                let metadata = Arc::new(CassResultData {
-                    paging_state: result.paging_state,
-                    col_specs: result.col_specs,
-                    tracing_id: result.tracing_id,
-                });
+                let metadata = Arc::new(CassResultData::from_result_payload(
+                    result.paging_state,
+                    result.col_specs,
+                    result.tracing_id,
+                ));
                 let cass_rows = create_cass_rows_from_rows(result.rows, &metadata);
                 let cass_result = Arc::new(CassResult {
                     rows: cass_rows,
@@ -333,9 +329,9 @@ fn create_cass_rows_from_rows(
 fn create_cass_row_columns(row: Row, metadata: &Arc<CassResultData>) -> Vec<CassValue> {
     row.columns
         .into_iter()
-        .zip(metadata.col_specs.iter())
-        .map(|(val, col)| {
-            let column_type = Arc::new(get_column_type(&col.typ));
+        .zip(metadata.col_data_types.iter())
+        .map(|(val, col_data_type)| {
+            let column_type = Arc::clone(col_data_type);
             CassValue {
                 value: val.map(|col_val| get_column_value(col_val, &column_type)),
                 value_type: column_type,