alamb commented on code in PR #8414:
URL: https://github.com/apache/arrow-datafusion/pull/8414#discussion_r1416232766
##########
datafusion/common/src/scalar.rs:
##########
@@ -1368,103 +1366,36 @@ impl ScalarValue {
}};
}
- macro_rules! build_array_list_primitive {
- ($ARRAY_TY:ident, $SCALAR_TY:ident, $NATIVE_TYPE:ident,
$LIST_TY:ident, $SCALAR_LIST:pat) => {{
- Ok::<ArrayRef,
DataFusionError>(Arc::new($LIST_TY::from_iter_primitive::<$ARRAY_TY, _, _>(
- scalars.into_iter().map(|x| match x{
- ScalarValue::List(arr) if matches!(x, $SCALAR_LIST) =>
{
- // `ScalarValue::List` contains a single element
`ListArray`.
- let list_arr = as_list_array(&arr);
- if list_arr.is_null(0) {
- Ok(None)
- } else {
- let primitive_arr =
-
list_arr.values().as_primitive::<$ARRAY_TY>();
- Ok(Some(
-
primitive_arr.into_iter().collect::<Vec<Option<_>>>(),
- ))
- }
- }
- ScalarValue::LargeList(arr) if matches!(x,
$SCALAR_LIST) =>{
- // `ScalarValue::List` contains a single element
`ListArray`.
- let list_arr = as_large_list_array(&arr);
- if list_arr.is_null(0) {
- Ok(None)
- } else {
- let primitive_arr =
-
list_arr.values().as_primitive::<$ARRAY_TY>();
- Ok(Some(
-
primitive_arr.into_iter().collect::<Vec<Option<_>>>(),
- ))
- }
- }
- sv => _internal_err!(
- "Inconsistent types in ScalarValue::iter_to_array.
\
- Expected {:?}, got {:?}",
- data_type, sv
- ),
- })
- .collect::<Result<Vec<_>>>()?,
- )))
- }};
- }
-
- macro_rules! build_array_list_string {
- ($BUILDER:ident,
$STRING_ARRAY:ident,$LIST_BUILDER:ident,$SCALAR_LIST:pat) => {{
- let mut builder = $LIST_BUILDER::new($BUILDER::new());
- for scalar in scalars.into_iter() {
- match scalar {
- ScalarValue::List(arr) if matches!(scalar,
$SCALAR_LIST) => {
- // `ScalarValue::List` contains a single element
`ListArray`.
- let list_arr = as_list_array(&arr);
-
- if list_arr.is_null(0) {
- builder.append(false);
- continue;
- }
-
- let string_arr = $STRING_ARRAY(list_arr.values());
-
- for v in string_arr.iter() {
- if let Some(v) = v {
- builder.values().append_value(v);
- } else {
- builder.values().append_null();
- }
- }
- builder.append(true);
- }
- ScalarValue::LargeList(arr) if matches!(scalar,
$SCALAR_LIST) => {
- // `ScalarValue::List` contains a single element
`ListArray`.
- let list_arr = as_large_list_array(&arr);
-
- if list_arr.is_null(0) {
- builder.append(false);
- continue;
- }
-
- let string_arr = $STRING_ARRAY(list_arr.values());
-
- for v in string_arr.iter() {
- if let Some(v) = v {
- builder.values().append_value(v);
- } else {
- builder.values().append_null();
- }
- }
- builder.append(true);
- }
- sv => {
- return _internal_err!(
- "Inconsistent types in
ScalarValue::iter_to_array. \
- Expected List, got {:?}",
- sv
- )
- }
- }
+ fn build_list_array(
+ scalars: impl IntoIterator<Item = ScalarValue>,
+ ) -> Result<ArrayRef> {
+ let arrays = scalars
Review Comment:
Is it necessary to make a 1 element array for each ScalarValue?
I wonder if you could make the values array directly like
```rust
let values = ScalarValue::iter_to_array(scalars);
// and then make a single offset of 0
```
🤔
##########
datafusion/common/src/scalar.rs:
##########
@@ -1368,103 +1366,36 @@ impl ScalarValue {
}};
}
- macro_rules! build_array_list_primitive {
- ($ARRAY_TY:ident, $SCALAR_TY:ident, $NATIVE_TYPE:ident,
$LIST_TY:ident, $SCALAR_LIST:pat) => {{
- Ok::<ArrayRef,
DataFusionError>(Arc::new($LIST_TY::from_iter_primitive::<$ARRAY_TY, _, _>(
- scalars.into_iter().map(|x| match x{
- ScalarValue::List(arr) if matches!(x, $SCALAR_LIST) =>
{
- // `ScalarValue::List` contains a single element
`ListArray`.
- let list_arr = as_list_array(&arr);
- if list_arr.is_null(0) {
- Ok(None)
- } else {
- let primitive_arr =
-
list_arr.values().as_primitive::<$ARRAY_TY>();
- Ok(Some(
-
primitive_arr.into_iter().collect::<Vec<Option<_>>>(),
- ))
- }
- }
- ScalarValue::LargeList(arr) if matches!(x,
$SCALAR_LIST) =>{
- // `ScalarValue::List` contains a single element
`ListArray`.
- let list_arr = as_large_list_array(&arr);
- if list_arr.is_null(0) {
- Ok(None)
- } else {
- let primitive_arr =
-
list_arr.values().as_primitive::<$ARRAY_TY>();
- Ok(Some(
-
primitive_arr.into_iter().collect::<Vec<Option<_>>>(),
- ))
- }
- }
- sv => _internal_err!(
- "Inconsistent types in ScalarValue::iter_to_array.
\
- Expected {:?}, got {:?}",
- data_type, sv
- ),
- })
- .collect::<Result<Vec<_>>>()?,
- )))
- }};
- }
-
- macro_rules! build_array_list_string {
- ($BUILDER:ident,
$STRING_ARRAY:ident,$LIST_BUILDER:ident,$SCALAR_LIST:pat) => {{
- let mut builder = $LIST_BUILDER::new($BUILDER::new());
- for scalar in scalars.into_iter() {
- match scalar {
- ScalarValue::List(arr) if matches!(scalar,
$SCALAR_LIST) => {
- // `ScalarValue::List` contains a single element
`ListArray`.
- let list_arr = as_list_array(&arr);
-
- if list_arr.is_null(0) {
- builder.append(false);
- continue;
- }
-
- let string_arr = $STRING_ARRAY(list_arr.values());
-
- for v in string_arr.iter() {
- if let Some(v) = v {
- builder.values().append_value(v);
- } else {
- builder.values().append_null();
- }
- }
- builder.append(true);
- }
- ScalarValue::LargeList(arr) if matches!(scalar,
$SCALAR_LIST) => {
- // `ScalarValue::List` contains a single element
`ListArray`.
- let list_arr = as_large_list_array(&arr);
-
- if list_arr.is_null(0) {
- builder.append(false);
- continue;
- }
-
- let string_arr = $STRING_ARRAY(list_arr.values());
-
- for v in string_arr.iter() {
- if let Some(v) = v {
- builder.values().append_value(v);
- } else {
- builder.values().append_null();
- }
- }
- builder.append(true);
- }
- sv => {
- return _internal_err!(
- "Inconsistent types in
ScalarValue::iter_to_array. \
- Expected List, got {:?}",
- sv
- )
- }
- }
+ fn build_list_array(
Review Comment:
😍 Amazing @jayzhan211
I believe as written, this function will always return `ListArray` (never
`LargeListArray`) so I think this probably needs to be templated on `<O:
OffsetSizeTrait>` -- the way @Weijun-H has done in
https://github.com/apache/arrow-datafusion/pull/8322 for example)
##########
datafusion/common/src/scalar.rs:
##########
@@ -1368,103 +1366,36 @@ impl ScalarValue {
}};
}
- macro_rules! build_array_list_primitive {
- ($ARRAY_TY:ident, $SCALAR_TY:ident, $NATIVE_TYPE:ident,
$LIST_TY:ident, $SCALAR_LIST:pat) => {{
- Ok::<ArrayRef,
DataFusionError>(Arc::new($LIST_TY::from_iter_primitive::<$ARRAY_TY, _, _>(
- scalars.into_iter().map(|x| match x{
- ScalarValue::List(arr) if matches!(x, $SCALAR_LIST) =>
{
- // `ScalarValue::List` contains a single element
`ListArray`.
- let list_arr = as_list_array(&arr);
- if list_arr.is_null(0) {
- Ok(None)
- } else {
- let primitive_arr =
-
list_arr.values().as_primitive::<$ARRAY_TY>();
- Ok(Some(
-
primitive_arr.into_iter().collect::<Vec<Option<_>>>(),
- ))
- }
- }
- ScalarValue::LargeList(arr) if matches!(x,
$SCALAR_LIST) =>{
- // `ScalarValue::List` contains a single element
`ListArray`.
- let list_arr = as_large_list_array(&arr);
- if list_arr.is_null(0) {
- Ok(None)
- } else {
- let primitive_arr =
-
list_arr.values().as_primitive::<$ARRAY_TY>();
- Ok(Some(
-
primitive_arr.into_iter().collect::<Vec<Option<_>>>(),
- ))
- }
- }
- sv => _internal_err!(
- "Inconsistent types in ScalarValue::iter_to_array.
\
- Expected {:?}, got {:?}",
- data_type, sv
- ),
- })
- .collect::<Result<Vec<_>>>()?,
- )))
- }};
- }
-
- macro_rules! build_array_list_string {
- ($BUILDER:ident,
$STRING_ARRAY:ident,$LIST_BUILDER:ident,$SCALAR_LIST:pat) => {{
- let mut builder = $LIST_BUILDER::new($BUILDER::new());
- for scalar in scalars.into_iter() {
- match scalar {
- ScalarValue::List(arr) if matches!(scalar,
$SCALAR_LIST) => {
- // `ScalarValue::List` contains a single element
`ListArray`.
- let list_arr = as_list_array(&arr);
-
- if list_arr.is_null(0) {
- builder.append(false);
- continue;
- }
-
- let string_arr = $STRING_ARRAY(list_arr.values());
-
- for v in string_arr.iter() {
- if let Some(v) = v {
- builder.values().append_value(v);
- } else {
- builder.values().append_null();
- }
- }
- builder.append(true);
- }
- ScalarValue::LargeList(arr) if matches!(scalar,
$SCALAR_LIST) => {
- // `ScalarValue::List` contains a single element
`ListArray`.
- let list_arr = as_large_list_array(&arr);
-
- if list_arr.is_null(0) {
- builder.append(false);
- continue;
- }
-
- let string_arr = $STRING_ARRAY(list_arr.values());
-
- for v in string_arr.iter() {
- if let Some(v) = v {
- builder.values().append_value(v);
- } else {
- builder.values().append_null();
- }
- }
- builder.append(true);
- }
- sv => {
- return _internal_err!(
- "Inconsistent types in
ScalarValue::iter_to_array. \
- Expected List, got {:?}",
- sv
- )
- }
- }
+ fn build_list_array(
Review Comment:
😍 Amazing @jayzhan211
I believe as written, this function will always return `ListArray` (never
`LargeListArray`) so I think this probably needs to be templated on `<O:
OffsetSizeTrait>` -- the way @Weijun-H has done in
https://github.com/apache/arrow-datafusion/pull/8322 for example)
##########
datafusion/common/src/scalar.rs:
##########
@@ -1368,103 +1366,36 @@ impl ScalarValue {
}};
}
- macro_rules! build_array_list_primitive {
- ($ARRAY_TY:ident, $SCALAR_TY:ident, $NATIVE_TYPE:ident,
$LIST_TY:ident, $SCALAR_LIST:pat) => {{
- Ok::<ArrayRef,
DataFusionError>(Arc::new($LIST_TY::from_iter_primitive::<$ARRAY_TY, _, _>(
- scalars.into_iter().map(|x| match x{
- ScalarValue::List(arr) if matches!(x, $SCALAR_LIST) =>
{
- // `ScalarValue::List` contains a single element
`ListArray`.
- let list_arr = as_list_array(&arr);
- if list_arr.is_null(0) {
- Ok(None)
- } else {
- let primitive_arr =
-
list_arr.values().as_primitive::<$ARRAY_TY>();
- Ok(Some(
-
primitive_arr.into_iter().collect::<Vec<Option<_>>>(),
- ))
- }
- }
- ScalarValue::LargeList(arr) if matches!(x,
$SCALAR_LIST) =>{
- // `ScalarValue::List` contains a single element
`ListArray`.
- let list_arr = as_large_list_array(&arr);
- if list_arr.is_null(0) {
- Ok(None)
- } else {
- let primitive_arr =
-
list_arr.values().as_primitive::<$ARRAY_TY>();
- Ok(Some(
-
primitive_arr.into_iter().collect::<Vec<Option<_>>>(),
- ))
- }
- }
- sv => _internal_err!(
- "Inconsistent types in ScalarValue::iter_to_array.
\
- Expected {:?}, got {:?}",
- data_type, sv
- ),
- })
- .collect::<Result<Vec<_>>>()?,
- )))
- }};
- }
-
- macro_rules! build_array_list_string {
- ($BUILDER:ident,
$STRING_ARRAY:ident,$LIST_BUILDER:ident,$SCALAR_LIST:pat) => {{
- let mut builder = $LIST_BUILDER::new($BUILDER::new());
- for scalar in scalars.into_iter() {
- match scalar {
- ScalarValue::List(arr) if matches!(scalar,
$SCALAR_LIST) => {
- // `ScalarValue::List` contains a single element
`ListArray`.
- let list_arr = as_list_array(&arr);
-
- if list_arr.is_null(0) {
- builder.append(false);
- continue;
- }
-
- let string_arr = $STRING_ARRAY(list_arr.values());
-
- for v in string_arr.iter() {
- if let Some(v) = v {
- builder.values().append_value(v);
- } else {
- builder.values().append_null();
- }
- }
- builder.append(true);
- }
- ScalarValue::LargeList(arr) if matches!(scalar,
$SCALAR_LIST) => {
- // `ScalarValue::List` contains a single element
`ListArray`.
- let list_arr = as_large_list_array(&arr);
-
- if list_arr.is_null(0) {
- builder.append(false);
- continue;
- }
-
- let string_arr = $STRING_ARRAY(list_arr.values());
-
- for v in string_arr.iter() {
- if let Some(v) = v {
- builder.values().append_value(v);
- } else {
- builder.values().append_null();
- }
- }
- builder.append(true);
- }
- sv => {
- return _internal_err!(
- "Inconsistent types in
ScalarValue::iter_to_array. \
- Expected List, got {:?}",
- sv
- )
- }
- }
+ fn build_list_array(
+ scalars: impl IntoIterator<Item = ScalarValue>,
+ ) -> Result<ArrayRef> {
+ let arrays = scalars
Review Comment:
Is it necessary to make a 1 element array for each ScalarValue?
I wonder if you could make the values array directly like
```rust
let values = ScalarValue::iter_to_array(scalars);
// and then make a single offset of 0
```
🤔
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]