clintropolis commented on code in PR #16516: URL: https://github.com/apache/druid/pull/16516#discussion_r1623144222
########## docs/release-info/migr-mvd-array.md: ########## @@ -0,0 +1,68 @@ +--- +id: migr-mvd-array +title: "Migration guide: MVDs to arrays" +sidebar_label: MVDs to arrays +--- + +<!-- + ~ Licensed to the Apache Software Foundation (ASF) under one + ~ or more contributor license agreements. See the NOTICE file + ~ distributed with this work for additional information + ~ regarding copyright ownership. The ASF licenses this file + ~ to you under the Apache License, Version 2.0 (the + ~ "License"); you may not use this file except in compliance + ~ with the License. You may obtain a copy of the License at + ~ + ~ http://www.apache.org/licenses/LICENSE-2.0 + ~ + ~ Unless required by applicable law or agreed to in writing, + ~ software distributed under the License is distributed on an + ~ "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY + ~ KIND, either express or implied. See the License for the + ~ specific language governing permissions and limitations + ~ under the License. +--> + + +Druid now supports SQL-compliant [arrays](../querying/arrays.md), and we recommend that people use arrays over [multi-value dimensions](../querying/multi-value-dimensions.md) (MVDs) whenever possible. +Use arrays for new projects and complex use cases involving multiple data types. Use MVDs for specific use cases, such as arrays of strings. + +## Comparison between arrays and MVDs + +The following table compares the general behavior between arrays and MVDs. +For specific query differences between arrays and MVDs, see [Query differences between arrays and MVDs](#query-differences-between-arrays-and-mvds). + +| | Arrays| Multi-value dimensions (MVDs) | +|---|---|---| +| Data types | Supports VARCHAR, BIGINT, and DOUBLE types (ARRAY<STRING\>, ARRAY<LONG\>, ARRAY<DOUBLE\>) | Only supports arrays of strings (VARCHAR) | +| SQL compliance | Behaves like standard SQL arrays with SQL-compliant behavior | Does not behave like standard SQL arrays; requires special SQL functions | +| Ingestion | <ul><li>Can be ingested using native batch or streaming ingestion methods</li><li>JSON arrays are ingested as Druid arrays</li><li>Managed through the query context parameter `arrayIngestMode` in SQL-based ingestion (supported options: `array`, `mvd`, `none`). Note that if you set this mode to `none`, Druid raises an exception if you try to store any type of array.</li></ul> | <ul><li>Typically ingested from fields with an array of values</li><li>JSON arrays are ingested as multi-value dimensions</li><li>Managed using functions like [ARRAY_TO_MV](../querying/sql-functions.md#array_to_mv) in SQL-based ingestion</li></ul> | +| Filtering and grouping | <ul><li>Filters and groupings match the entire array value</li><li>Can be used as GROUP BY keys, grouping based on the entire array value</li></ul> | <ul><li>Filters match any value within the array</li><li>Grouping generates a group for each individual value, similar to an implicit UNNEST</li></ul> | +| Conversion | Convert an MVD to an array using [MV_TO_ARRAY](../querying/sql-multivalue-string-functions.md) | Convert an array to an MVD using [ARRAY_TO_MV](../querying/sql-functions.md#array_to_mv) | + +### Query differences between arrays and MVDs Review Comment: i worry this section is a bit more confusing than it is helpful, there are a lot of examples of behavior and differences between arrays and mvds in https://github.com/apache/druid/blob/master/docs/querying/arrays.md and https://github.com/apache/druid/blob/master/docs/querying/multi-value-dimensions.md currently this table isn't super clear if it is talking about native or SQL filters, and in some cases maybe is talking about both? (like array_overlap isn't a native filter, but there is a native `arrayContainsElement` which when combined with a native `or` filter can construct ARRAY_OVERLAP). I fear this content is too deep to summarize into this table and we might be better off delegating into the in depth docs which have examples ########## docs/release-info/migr-mvd-array.md: ########## @@ -0,0 +1,68 @@ +--- +id: migr-mvd-array +title: "Migration guide: MVDs to arrays" +sidebar_label: MVDs to arrays +--- + +<!-- + ~ Licensed to the Apache Software Foundation (ASF) under one + ~ or more contributor license agreements. See the NOTICE file + ~ distributed with this work for additional information + ~ regarding copyright ownership. The ASF licenses this file + ~ to you under the Apache License, Version 2.0 (the + ~ "License"); you may not use this file except in compliance + ~ with the License. You may obtain a copy of the License at + ~ + ~ http://www.apache.org/licenses/LICENSE-2.0 + ~ + ~ Unless required by applicable law or agreed to in writing, + ~ software distributed under the License is distributed on an + ~ "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY + ~ KIND, either express or implied. See the License for the + ~ specific language governing permissions and limitations + ~ under the License. +--> + + +Druid now supports SQL-compliant [arrays](../querying/arrays.md), and we recommend that people use arrays over [multi-value dimensions](../querying/multi-value-dimensions.md) (MVDs) whenever possible. +Use arrays for new projects and complex use cases involving multiple data types. Use MVDs for specific use cases, such as arrays of strings. + +## Comparison between arrays and MVDs + +The following table compares the general behavior between arrays and MVDs. +For specific query differences between arrays and MVDs, see [Query differences between arrays and MVDs](#query-differences-between-arrays-and-mvds). + +| | Arrays| Multi-value dimensions (MVDs) | +|---|---|---| +| Data types | Supports VARCHAR, BIGINT, and DOUBLE types (ARRAY<STRING\>, ARRAY<LONG\>, ARRAY<DOUBLE\>) | Only supports arrays of strings (VARCHAR) | +| SQL compliance | Behaves like standard SQL arrays with SQL-compliant behavior | Does not behave like standard SQL arrays; requires special SQL functions | +| Ingestion | <ul><li>Can be ingested using native batch or streaming ingestion methods</li><li>JSON arrays are ingested as Druid arrays</li><li>Managed through the query context parameter `arrayIngestMode` in SQL-based ingestion (supported options: `array`, `mvd`, `none`). Note that if you set this mode to `none`, Druid raises an exception if you try to store any type of array.</li></ul> | <ul><li>Typically ingested from fields with an array of values</li><li>JSON arrays are ingested as multi-value dimensions</li><li>Managed using functions like [ARRAY_TO_MV](../querying/sql-functions.md#array_to_mv) in SQL-based ingestion</li></ul> | +| Filtering and grouping | <ul><li>Filters and groupings match the entire array value</li><li>Can be used as GROUP BY keys, grouping based on the entire array value</li></ul> | <ul><li>Filters match any value within the array</li><li>Grouping generates a group for each individual value, similar to an implicit UNNEST</li></ul> | +| Conversion | Convert an MVD to an array using [MV_TO_ARRAY](../querying/sql-multivalue-string-functions.md) | Convert an array to an MVD using [ARRAY_TO_MV](../querying/sql-functions.md#array_to_mv) | + +### Query differences between arrays and MVDs + +Querying an array column returns different results than when querying a MVD column depending on the query type. Review the following table on the differences. Adjust your applications to handle the new output shapes. + +| Query type | Array | MVD | +|---|---|---| +| Equality filter | Matches the entire array value | Matches if any value within the array matches the filter | +| Null filter | Matches rows where the entire array value is null | Matches rows where the array is empty (considered as null) but does not match arrays with empty (`“”`) values | +| Range filter | Use [ARRAY_OVERLAP](../querying/sql-functions.md#array_overlap) | Not directly supported | +| Contains filter | Use [ARRAY_CONTAINS](../querying/sql-functions.md#array_contains)| Use WHERE filter | +| Logical expression filters | Behaves like standard ANSI SQL on the entire array value, such as AND, OR, NOT. For example, `WHERE arrayLong = ARRAY[1,2,3] OR arrayLong = ARRAY[4,5,6]` | Matches a row if any value within the array matches the logical condition. For example, `WHERE tags = 't1' OR tags = 't3'` | +| Column comparison filter | Use [ARRAY_OVERLAP](../querying/sql-functions.md#array_overlap) | Matches when the dimensions have any overlapping values. For example, `WHERE tags IN ('t1', 't2')` | +| Behavior with SQL constructs | Follows standard SQL behavior with array functions like [ARRAY_CONTAINS](../querying/sql-functions.md#array_contains), [ARRAY_OVERLAP](../querying/sql-functions.md#array_overlap) | Requires special SQL functions like [MV_FILTER_ONLY](../querying/sql-functions.md#mv_filter_none), [MV_FILTER_NONE](../querying/sql-functions.md#mv_filter_only) for precise filtering | +| Group by entire array | Groups the entire array as a single value | Not supported | +| Group by individual values | Use [UNNEST](../querying/sql.md#unnest) to group by individual array elements | Automatically unnests groups by each individual value in the array | + +## How to ingest data as arrays + Review Comment: again i think maybe should just refer to the other external docs which include examples of how to ingest both native and sql for both arrays and mvds ########## docs/release-info/migr-mvd-array.md: ########## @@ -0,0 +1,68 @@ +--- +id: migr-mvd-array +title: "Migration guide: MVDs to arrays" +sidebar_label: MVDs to arrays +--- + +<!-- + ~ Licensed to the Apache Software Foundation (ASF) under one + ~ or more contributor license agreements. See the NOTICE file + ~ distributed with this work for additional information + ~ regarding copyright ownership. The ASF licenses this file + ~ to you under the Apache License, Version 2.0 (the + ~ "License"); you may not use this file except in compliance + ~ with the License. You may obtain a copy of the License at + ~ + ~ http://www.apache.org/licenses/LICENSE-2.0 + ~ + ~ Unless required by applicable law or agreed to in writing, + ~ software distributed under the License is distributed on an + ~ "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY + ~ KIND, either express or implied. See the License for the + ~ specific language governing permissions and limitations + ~ under the License. +--> + + +Druid now supports SQL-compliant [arrays](../querying/arrays.md), and we recommend that people use arrays over [multi-value dimensions](../querying/multi-value-dimensions.md) (MVDs) whenever possible. +Use arrays for new projects and complex use cases involving multiple data types. Use MVDs for specific use cases, such as arrays of strings. Review Comment: >Use MVDs for specific use cases, such as arrays of strings. This seems off, the use case for MVDs is really when the user doesn't care about the complete array of values (including ordering of the multiple values within the row) as a value itself, rather are only concerned with operating directly on the elements as if they were regular strings. ########## docs/release-info/migr-mvd-array.md: ########## @@ -0,0 +1,68 @@ +--- +id: migr-mvd-array +title: "Migration guide: MVDs to arrays" +sidebar_label: MVDs to arrays +--- + +<!-- + ~ Licensed to the Apache Software Foundation (ASF) under one + ~ or more contributor license agreements. See the NOTICE file + ~ distributed with this work for additional information + ~ regarding copyright ownership. The ASF licenses this file + ~ to you under the Apache License, Version 2.0 (the + ~ "License"); you may not use this file except in compliance + ~ with the License. You may obtain a copy of the License at + ~ + ~ http://www.apache.org/licenses/LICENSE-2.0 + ~ + ~ Unless required by applicable law or agreed to in writing, + ~ software distributed under the License is distributed on an + ~ "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY + ~ KIND, either express or implied. See the License for the + ~ specific language governing permissions and limitations + ~ under the License. +--> + + +Druid now supports SQL-compliant [arrays](../querying/arrays.md), and we recommend that people use arrays over [multi-value dimensions](../querying/multi-value-dimensions.md) (MVDs) whenever possible. +Use arrays for new projects and complex use cases involving multiple data types. Use MVDs for specific use cases, such as arrays of strings. + +## Comparison between arrays and MVDs + +The following table compares the general behavior between arrays and MVDs. +For specific query differences between arrays and MVDs, see [Query differences between arrays and MVDs](#query-differences-between-arrays-and-mvds). + +| | Arrays| Multi-value dimensions (MVDs) | +|---|---|---| +| Data types | Supports VARCHAR, BIGINT, and DOUBLE types (ARRAY<STRING\>, ARRAY<LONG\>, ARRAY<DOUBLE\>) | Only supports arrays of strings (VARCHAR) | +| SQL compliance | Behaves like standard SQL arrays with SQL-compliant behavior | Does not behave like standard SQL arrays; requires special SQL functions | +| Ingestion | <ul><li>Can be ingested using native batch or streaming ingestion methods</li><li>JSON arrays are ingested as Druid arrays</li><li>Managed through the query context parameter `arrayIngestMode` in SQL-based ingestion (supported options: `array`, `mvd`, `none`). Note that if you set this mode to `none`, Druid raises an exception if you try to store any type of array.</li></ul> | <ul><li>Typically ingested from fields with an array of values</li><li>JSON arrays are ingested as multi-value dimensions</li><li>Managed using functions like [ARRAY_TO_MV](../querying/sql-functions.md#array_to_mv) in SQL-based ingestion</li></ul> | +| Filtering and grouping | <ul><li>Filters and groupings match the entire array value</li><li>Can be used as GROUP BY keys, grouping based on the entire array value</li></ul> | <ul><li>Filters match any value within the array</li><li>Grouping generates a group for each individual value, similar to an implicit UNNEST</li></ul> | +| Conversion | Convert an MVD to an array using [MV_TO_ARRAY](../querying/sql-multivalue-string-functions.md) | Convert an array to an MVD using [ARRAY_TO_MV](../querying/sql-functions.md#array_to_mv) | + +### Query differences between arrays and MVDs + +Querying an array column returns different results than when querying a MVD column depending on the query type. Review the following table on the differences. Adjust your applications to handle the new output shapes. + +| Query type | Array | MVD | +|---|---|---| +| Equality filter | Matches the entire array value | Matches if any value within the array matches the filter | +| Null filter | Matches rows where the entire array value is null | Matches rows where the array is empty (considered as null) but does not match arrays with empty (`“”`) values | +| Range filter | Use [ARRAY_OVERLAP](../querying/sql-functions.md#array_overlap) | Not directly supported | Review Comment: arrays support range filters using comparison of the sort order of the whole array, e.g. `[1, 2, 3]` < `[1, 2, 4]` if i write `... WHERE arrayColumn < ARRAY[1, 2, 3]` or whatever. Array overlap is a different filter that checks if an array contains any of the elements of some other array. Multi-value dimensions also support range filters if matching as individual string values. ########## docs/release-info/migr-mvd-array.md: ########## @@ -0,0 +1,68 @@ +--- +id: migr-mvd-array +title: "Migration guide: MVDs to arrays" +sidebar_label: MVDs to arrays +--- + +<!-- + ~ Licensed to the Apache Software Foundation (ASF) under one + ~ or more contributor license agreements. See the NOTICE file + ~ distributed with this work for additional information + ~ regarding copyright ownership. The ASF licenses this file + ~ to you under the Apache License, Version 2.0 (the + ~ "License"); you may not use this file except in compliance + ~ with the License. You may obtain a copy of the License at + ~ + ~ http://www.apache.org/licenses/LICENSE-2.0 + ~ + ~ Unless required by applicable law or agreed to in writing, + ~ software distributed under the License is distributed on an + ~ "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY + ~ KIND, either express or implied. See the License for the + ~ specific language governing permissions and limitations + ~ under the License. +--> + + +Druid now supports SQL-compliant [arrays](../querying/arrays.md), and we recommend that people use arrays over [multi-value dimensions](../querying/multi-value-dimensions.md) (MVDs) whenever possible. +Use arrays for new projects and complex use cases involving multiple data types. Use MVDs for specific use cases, such as arrays of strings. + +## Comparison between arrays and MVDs + +The following table compares the general behavior between arrays and MVDs. +For specific query differences between arrays and MVDs, see [Query differences between arrays and MVDs](#query-differences-between-arrays-and-mvds). + +| | Arrays| Multi-value dimensions (MVDs) | +|---|---|---| +| Data types | Supports VARCHAR, BIGINT, and DOUBLE types (ARRAY<STRING\>, ARRAY<LONG\>, ARRAY<DOUBLE\>) | Only supports arrays of strings (VARCHAR) | +| SQL compliance | Behaves like standard SQL arrays with SQL-compliant behavior | Does not behave like standard SQL arrays; requires special SQL functions | +| Ingestion | <ul><li>Can be ingested using native batch or streaming ingestion methods</li><li>JSON arrays are ingested as Druid arrays</li><li>Managed through the query context parameter `arrayIngestMode` in SQL-based ingestion (supported options: `array`, `mvd`, `none`). Note that if you set this mode to `none`, Druid raises an exception if you try to store any type of array.</li></ul> | <ul><li>Typically ingested from fields with an array of values</li><li>JSON arrays are ingested as multi-value dimensions</li><li>Managed using functions like [ARRAY_TO_MV](../querying/sql-functions.md#array_to_mv) in SQL-based ingestion</li></ul> | +| Filtering and grouping | <ul><li>Filters and groupings match the entire array value</li><li>Can be used as GROUP BY keys, grouping based on the entire array value</li></ul> | <ul><li>Filters match any value within the array</li><li>Grouping generates a group for each individual value, similar to an implicit UNNEST</li></ul> | +| Conversion | Convert an MVD to an array using [MV_TO_ARRAY](../querying/sql-multivalue-string-functions.md) | Convert an array to an MVD using [ARRAY_TO_MV](../querying/sql-functions.md#array_to_mv) | + +### Query differences between arrays and MVDs + +Querying an array column returns different results than when querying a MVD column depending on the query type. Review the following table on the differences. Adjust your applications to handle the new output shapes. + +| Query type | Array | MVD | +|---|---|---| +| Equality filter | Matches the entire array value | Matches if any value within the array matches the filter | +| Null filter | Matches rows where the entire array value is null | Matches rows where the array is empty (considered as null) but does not match arrays with empty (`“”`) values | +| Range filter | Use [ARRAY_OVERLAP](../querying/sql-functions.md#array_overlap) | Not directly supported | +| Contains filter | Use [ARRAY_CONTAINS](../querying/sql-functions.md#array_contains)| Use WHERE filter | +| Logical expression filters | Behaves like standard ANSI SQL on the entire array value, such as AND, OR, NOT. For example, `WHERE arrayLong = ARRAY[1,2,3] OR arrayLong = ARRAY[4,5,6]` | Matches a row if any value within the array matches the logical condition. For example, `WHERE tags = 't1' OR tags = 't3'` | +| Column comparison filter | Use [ARRAY_OVERLAP](../querying/sql-functions.md#array_overlap) | Matches when the dimensions have any overlapping values. For example, `WHERE tags IN ('t1', 't2')` | Review Comment: the native column comparison filter itself is very strange (and not used by sql at all), so i think we should drop it from here. It converts everything into a string array for comparison regardless of type, so it works more or less the same for both mvds and arrays, ish, but again is very weird and i feel like only makes things more confusing ########## docs/release-info/migr-mvd-array.md: ########## @@ -0,0 +1,68 @@ +--- +id: migr-mvd-array +title: "Migration guide: MVDs to arrays" +sidebar_label: MVDs to arrays +--- + +<!-- + ~ Licensed to the Apache Software Foundation (ASF) under one + ~ or more contributor license agreements. See the NOTICE file + ~ distributed with this work for additional information + ~ regarding copyright ownership. The ASF licenses this file + ~ to you under the Apache License, Version 2.0 (the + ~ "License"); you may not use this file except in compliance + ~ with the License. You may obtain a copy of the License at + ~ + ~ http://www.apache.org/licenses/LICENSE-2.0 + ~ + ~ Unless required by applicable law or agreed to in writing, + ~ software distributed under the License is distributed on an + ~ "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY + ~ KIND, either express or implied. See the License for the + ~ specific language governing permissions and limitations + ~ under the License. +--> + + +Druid now supports SQL-compliant [arrays](../querying/arrays.md), and we recommend that people use arrays over [multi-value dimensions](../querying/multi-value-dimensions.md) (MVDs) whenever possible. +Use arrays for new projects and complex use cases involving multiple data types. Use MVDs for specific use cases, such as arrays of strings. + +## Comparison between arrays and MVDs + +The following table compares the general behavior between arrays and MVDs. +For specific query differences between arrays and MVDs, see [Query differences between arrays and MVDs](#query-differences-between-arrays-and-mvds). + +| | Arrays| Multi-value dimensions (MVDs) | +|---|---|---| +| Data types | Supports VARCHAR, BIGINT, and DOUBLE types (ARRAY<STRING\>, ARRAY<LONG\>, ARRAY<DOUBLE\>) | Only supports arrays of strings (VARCHAR) | +| SQL compliance | Behaves like standard SQL arrays with SQL-compliant behavior | Does not behave like standard SQL arrays; requires special SQL functions | +| Ingestion | <ul><li>Can be ingested using native batch or streaming ingestion methods</li><li>JSON arrays are ingested as Druid arrays</li><li>Managed through the query context parameter `arrayIngestMode` in SQL-based ingestion (supported options: `array`, `mvd`, `none`). Note that if you set this mode to `none`, Druid raises an exception if you try to store any type of array.</li></ul> | <ul><li>Typically ingested from fields with an array of values</li><li>JSON arrays are ingested as multi-value dimensions</li><li>Managed using functions like [ARRAY_TO_MV](../querying/sql-functions.md#array_to_mv) in SQL-based ingestion</li></ul> | +| Filtering and grouping | <ul><li>Filters and groupings match the entire array value</li><li>Can be used as GROUP BY keys, grouping based on the entire array value</li></ul> | <ul><li>Filters match any value within the array</li><li>Grouping generates a group for each individual value, similar to an implicit UNNEST</li></ul> | +| Conversion | Convert an MVD to an array using [MV_TO_ARRAY](../querying/sql-multivalue-string-functions.md) | Convert an array to an MVD using [ARRAY_TO_MV](../querying/sql-functions.md#array_to_mv) | + +### Query differences between arrays and MVDs + +Querying an array column returns different results than when querying a MVD column depending on the query type. Review the following table on the differences. Adjust your applications to handle the new output shapes. + +| Query type | Array | MVD | +|---|---|---| +| Equality filter | Matches the entire array value | Matches if any value within the array matches the filter | Review Comment: I think it isn't clear enough that the biggest difference between arrays and mvds is how you interact with them in SQL. like array types you treat as SQL arrays, while mvds you treat as SQL VARCHAR. Like if i have two columns with the same data, some array `['a', 'b', 'c']`, then for the array type the only thing that matches that row is`WHERE array = ARRAY['a', 'b', 'c']`, while any of `WHERE mvd = 'a'`, `WHERE mvd = 'b'`, `WHERE mvd = 'c'` would match the row. ########## docs/release-info/migr-mvd-array.md: ########## @@ -0,0 +1,68 @@ +--- +id: migr-mvd-array +title: "Migration guide: MVDs to arrays" +sidebar_label: MVDs to arrays +--- + +<!-- + ~ Licensed to the Apache Software Foundation (ASF) under one + ~ or more contributor license agreements. See the NOTICE file + ~ distributed with this work for additional information + ~ regarding copyright ownership. The ASF licenses this file + ~ to you under the Apache License, Version 2.0 (the + ~ "License"); you may not use this file except in compliance + ~ with the License. You may obtain a copy of the License at + ~ + ~ http://www.apache.org/licenses/LICENSE-2.0 + ~ + ~ Unless required by applicable law or agreed to in writing, + ~ software distributed under the License is distributed on an + ~ "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY + ~ KIND, either express or implied. See the License for the + ~ specific language governing permissions and limitations + ~ under the License. +--> + + +Druid now supports SQL-compliant [arrays](../querying/arrays.md), and we recommend that people use arrays over [multi-value dimensions](../querying/multi-value-dimensions.md) (MVDs) whenever possible. +Use arrays for new projects and complex use cases involving multiple data types. Use MVDs for specific use cases, such as arrays of strings. + +## Comparison between arrays and MVDs + +The following table compares the general behavior between arrays and MVDs. +For specific query differences between arrays and MVDs, see [Query differences between arrays and MVDs](#query-differences-between-arrays-and-mvds). + +| | Arrays| Multi-value dimensions (MVDs) | +|---|---|---| +| Data types | Supports VARCHAR, BIGINT, and DOUBLE types (ARRAY<STRING\>, ARRAY<LONG\>, ARRAY<DOUBLE\>) | Only supports arrays of strings (VARCHAR) | +| SQL compliance | Behaves like standard SQL arrays with SQL-compliant behavior | Does not behave like standard SQL arrays; requires special SQL functions | +| Ingestion | <ul><li>Can be ingested using native batch or streaming ingestion methods</li><li>JSON arrays are ingested as Druid arrays</li><li>Managed through the query context parameter `arrayIngestMode` in SQL-based ingestion (supported options: `array`, `mvd`, `none`). Note that if you set this mode to `none`, Druid raises an exception if you try to store any type of array.</li></ul> | <ul><li>Typically ingested from fields with an array of values</li><li>JSON arrays are ingested as multi-value dimensions</li><li>Managed using functions like [ARRAY_TO_MV](../querying/sql-functions.md#array_to_mv) in SQL-based ingestion</li></ul> | +| Filtering and grouping | <ul><li>Filters and groupings match the entire array value</li><li>Can be used as GROUP BY keys, grouping based on the entire array value</li></ul> | <ul><li>Filters match any value within the array</li><li>Grouping generates a group for each individual value, similar to an implicit UNNEST</li></ul> | +| Conversion | Convert an MVD to an array using [MV_TO_ARRAY](../querying/sql-multivalue-string-functions.md) | Convert an array to an MVD using [ARRAY_TO_MV](../querying/sql-functions.md#array_to_mv) | + +### Query differences between arrays and MVDs + +Querying an array column returns different results than when querying a MVD column depending on the query type. Review the following table on the differences. Adjust your applications to handle the new output shapes. + +| Query type | Array | MVD | +|---|---|---| +| Equality filter | Matches the entire array value | Matches if any value within the array matches the filter | +| Null filter | Matches rows where the entire array value is null | Matches rows where the array is empty (considered as null) but does not match arrays with empty (`“”`) values | +| Range filter | Use [ARRAY_OVERLAP](../querying/sql-functions.md#array_overlap) | Not directly supported | +| Contains filter | Use [ARRAY_CONTAINS](../querying/sql-functions.md#array_contains)| Use WHERE filter | +| Logical expression filters | Behaves like standard ANSI SQL on the entire array value, such as AND, OR, NOT. For example, `WHERE arrayLong = ARRAY[1,2,3] OR arrayLong = ARRAY[4,5,6]` | Matches a row if any value within the array matches the logical condition. For example, `WHERE tags = 't1' OR tags = 't3'` | +| Column comparison filter | Use [ARRAY_OVERLAP](../querying/sql-functions.md#array_overlap) | Matches when the dimensions have any overlapping values. For example, `WHERE tags IN ('t1', 't2')` | +| Behavior with SQL constructs | Follows standard SQL behavior with array functions like [ARRAY_CONTAINS](../querying/sql-functions.md#array_contains), [ARRAY_OVERLAP](../querying/sql-functions.md#array_overlap) | Requires special SQL functions like [MV_FILTER_ONLY](../querying/sql-functions.md#mv_filter_none), [MV_FILTER_NONE](../querying/sql-functions.md#mv_filter_only) for precise filtering | +| Group by entire array | Groups the entire array as a single value | Not supported | Review Comment: you can use `mv_to_array` to group mvds as arrays ########## docs/release-info/migr-mvd-array.md: ########## @@ -0,0 +1,68 @@ +--- +id: migr-mvd-array +title: "Migration guide: MVDs to arrays" +sidebar_label: MVDs to arrays +--- + +<!-- + ~ Licensed to the Apache Software Foundation (ASF) under one + ~ or more contributor license agreements. See the NOTICE file + ~ distributed with this work for additional information + ~ regarding copyright ownership. The ASF licenses this file + ~ to you under the Apache License, Version 2.0 (the + ~ "License"); you may not use this file except in compliance + ~ with the License. You may obtain a copy of the License at + ~ + ~ http://www.apache.org/licenses/LICENSE-2.0 + ~ + ~ Unless required by applicable law or agreed to in writing, + ~ software distributed under the License is distributed on an + ~ "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY + ~ KIND, either express or implied. See the License for the + ~ specific language governing permissions and limitations + ~ under the License. +--> + + +Druid now supports SQL-compliant [arrays](../querying/arrays.md), and we recommend that people use arrays over [multi-value dimensions](../querying/multi-value-dimensions.md) (MVDs) whenever possible. +Use arrays for new projects and complex use cases involving multiple data types. Use MVDs for specific use cases, such as arrays of strings. + +## Comparison between arrays and MVDs + +The following table compares the general behavior between arrays and MVDs. +For specific query differences between arrays and MVDs, see [Query differences between arrays and MVDs](#query-differences-between-arrays-and-mvds). + +| | Arrays| Multi-value dimensions (MVDs) | +|---|---|---| +| Data types | Supports VARCHAR, BIGINT, and DOUBLE types (ARRAY<STRING\>, ARRAY<LONG\>, ARRAY<DOUBLE\>) | Only supports arrays of strings (VARCHAR) | +| SQL compliance | Behaves like standard SQL arrays with SQL-compliant behavior | Does not behave like standard SQL arrays; requires special SQL functions | +| Ingestion | <ul><li>Can be ingested using native batch or streaming ingestion methods</li><li>JSON arrays are ingested as Druid arrays</li><li>Managed through the query context parameter `arrayIngestMode` in SQL-based ingestion (supported options: `array`, `mvd`, `none`). Note that if you set this mode to `none`, Druid raises an exception if you try to store any type of array.</li></ul> | <ul><li>Typically ingested from fields with an array of values</li><li>JSON arrays are ingested as multi-value dimensions</li><li>Managed using functions like [ARRAY_TO_MV](../querying/sql-functions.md#array_to_mv) in SQL-based ingestion</li></ul> | +| Filtering and grouping | <ul><li>Filters and groupings match the entire array value</li><li>Can be used as GROUP BY keys, grouping based on the entire array value</li></ul> | <ul><li>Filters match any value within the array</li><li>Grouping generates a group for each individual value, similar to an implicit UNNEST</li></ul> | +| Conversion | Convert an MVD to an array using [MV_TO_ARRAY](../querying/sql-multivalue-string-functions.md) | Convert an array to an MVD using [ARRAY_TO_MV](../querying/sql-functions.md#array_to_mv) | + +### Query differences between arrays and MVDs + +Querying an array column returns different results than when querying a MVD column depending on the query type. Review the following table on the differences. Adjust your applications to handle the new output shapes. + +| Query type | Array | MVD | +|---|---|---| +| Equality filter | Matches the entire array value | Matches if any value within the array matches the filter | +| Null filter | Matches rows where the entire array value is null | Matches rows where the array is empty (considered as null) but does not match arrays with empty (`“”`) values | +| Range filter | Use [ARRAY_OVERLAP](../querying/sql-functions.md#array_overlap) | Not directly supported | +| Contains filter | Use [ARRAY_CONTAINS](../querying/sql-functions.md#array_contains)| Use WHERE filter | +| Logical expression filters | Behaves like standard ANSI SQL on the entire array value, such as AND, OR, NOT. For example, `WHERE arrayLong = ARRAY[1,2,3] OR arrayLong = ARRAY[4,5,6]` | Matches a row if any value within the array matches the logical condition. For example, `WHERE tags = 't1' OR tags = 't3'` | Review Comment: this seems confusing and should be left out i think, since the logical filters operate on the results of the child filters rather than directly on arrays or mvds ########## docs/release-info/migr-mvd-array.md: ########## @@ -0,0 +1,68 @@ +--- +id: migr-mvd-array +title: "Migration guide: MVDs to arrays" +sidebar_label: MVDs to arrays +--- + +<!-- + ~ Licensed to the Apache Software Foundation (ASF) under one + ~ or more contributor license agreements. See the NOTICE file + ~ distributed with this work for additional information + ~ regarding copyright ownership. The ASF licenses this file + ~ to you under the Apache License, Version 2.0 (the + ~ "License"); you may not use this file except in compliance + ~ with the License. You may obtain a copy of the License at + ~ + ~ http://www.apache.org/licenses/LICENSE-2.0 + ~ + ~ Unless required by applicable law or agreed to in writing, + ~ software distributed under the License is distributed on an + ~ "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY + ~ KIND, either express or implied. See the License for the + ~ specific language governing permissions and limitations + ~ under the License. +--> + + +Druid now supports SQL-compliant [arrays](../querying/arrays.md), and we recommend that people use arrays over [multi-value dimensions](../querying/multi-value-dimensions.md) (MVDs) whenever possible. +Use arrays for new projects and complex use cases involving multiple data types. Use MVDs for specific use cases, such as arrays of strings. + +## Comparison between arrays and MVDs + +The following table compares the general behavior between arrays and MVDs. +For specific query differences between arrays and MVDs, see [Query differences between arrays and MVDs](#query-differences-between-arrays-and-mvds). + +| | Arrays| Multi-value dimensions (MVDs) | +|---|---|---| +| Data types | Supports VARCHAR, BIGINT, and DOUBLE types (ARRAY<STRING\>, ARRAY<LONG\>, ARRAY<DOUBLE\>) | Only supports arrays of strings (VARCHAR) | +| SQL compliance | Behaves like standard SQL arrays with SQL-compliant behavior | Does not behave like standard SQL arrays; requires special SQL functions | +| Ingestion | <ul><li>Can be ingested using native batch or streaming ingestion methods</li><li>JSON arrays are ingested as Druid arrays</li><li>Managed through the query context parameter `arrayIngestMode` in SQL-based ingestion (supported options: `array`, `mvd`, `none`). Note that if you set this mode to `none`, Druid raises an exception if you try to store any type of array.</li></ul> | <ul><li>Typically ingested from fields with an array of values</li><li>JSON arrays are ingested as multi-value dimensions</li><li>Managed using functions like [ARRAY_TO_MV](../querying/sql-functions.md#array_to_mv) in SQL-based ingestion</li></ul> | +| Filtering and grouping | <ul><li>Filters and groupings match the entire array value</li><li>Can be used as GROUP BY keys, grouping based on the entire array value</li></ul> | <ul><li>Filters match any value within the array</li><li>Grouping generates a group for each individual value, similar to an implicit UNNEST</li></ul> | +| Conversion | Convert an MVD to an array using [MV_TO_ARRAY](../querying/sql-multivalue-string-functions.md) | Convert an array to an MVD using [ARRAY_TO_MV](../querying/sql-functions.md#array_to_mv) | + +### Query differences between arrays and MVDs + +Querying an array column returns different results than when querying a MVD column depending on the query type. Review the following table on the differences. Adjust your applications to handle the new output shapes. + +| Query type | Array | MVD | +|---|---|---| +| Equality filter | Matches the entire array value | Matches if any value within the array matches the filter | +| Null filter | Matches rows where the entire array value is null | Matches rows where the array is empty (considered as null) but does not match arrays with empty (`“”`) values | +| Range filter | Use [ARRAY_OVERLAP](../querying/sql-functions.md#array_overlap) | Not directly supported | +| Contains filter | Use [ARRAY_CONTAINS](../querying/sql-functions.md#array_contains)| Use WHERE filter | Review Comment: is "Contains filter" referring to the native `arrayContainsElement` filter? Also "Use WHERE filter" isn't correct, perhaps you meant "Use equality filter" instead? ########## docs/release-info/migr-mvd-array.md: ########## @@ -0,0 +1,68 @@ +--- +id: migr-mvd-array +title: "Migration guide: MVDs to arrays" +sidebar_label: MVDs to arrays +--- + +<!-- + ~ Licensed to the Apache Software Foundation (ASF) under one + ~ or more contributor license agreements. See the NOTICE file + ~ distributed with this work for additional information + ~ regarding copyright ownership. The ASF licenses this file + ~ to you under the Apache License, Version 2.0 (the + ~ "License"); you may not use this file except in compliance + ~ with the License. You may obtain a copy of the License at + ~ + ~ http://www.apache.org/licenses/LICENSE-2.0 + ~ + ~ Unless required by applicable law or agreed to in writing, + ~ software distributed under the License is distributed on an + ~ "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY + ~ KIND, either express or implied. See the License for the + ~ specific language governing permissions and limitations + ~ under the License. +--> + + +Druid now supports SQL-compliant [arrays](../querying/arrays.md), and we recommend that people use arrays over [multi-value dimensions](../querying/multi-value-dimensions.md) (MVDs) whenever possible. +Use arrays for new projects and complex use cases involving multiple data types. Use MVDs for specific use cases, such as arrays of strings. + +## Comparison between arrays and MVDs + +The following table compares the general behavior between arrays and MVDs. +For specific query differences between arrays and MVDs, see [Query differences between arrays and MVDs](#query-differences-between-arrays-and-mvds). + +| | Arrays| Multi-value dimensions (MVDs) | +|---|---|---| +| Data types | Supports VARCHAR, BIGINT, and DOUBLE types (ARRAY<STRING\>, ARRAY<LONG\>, ARRAY<DOUBLE\>) | Only supports arrays of strings (VARCHAR) | +| SQL compliance | Behaves like standard SQL arrays with SQL-compliant behavior | Does not behave like standard SQL arrays; requires special SQL functions | +| Ingestion | <ul><li>Can be ingested using native batch or streaming ingestion methods</li><li>JSON arrays are ingested as Druid arrays</li><li>Managed through the query context parameter `arrayIngestMode` in SQL-based ingestion (supported options: `array`, `mvd`, `none`). Note that if you set this mode to `none`, Druid raises an exception if you try to store any type of array.</li></ul> | <ul><li>Typically ingested from fields with an array of values</li><li>JSON arrays are ingested as multi-value dimensions</li><li>Managed using functions like [ARRAY_TO_MV](../querying/sql-functions.md#array_to_mv) in SQL-based ingestion</li></ul> | Review Comment: yea, it seems a bit confusing. this seems tricky to follow without examples -- 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: commits-unsubscr...@druid.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: commits-unsubscr...@druid.apache.org For additional commands, e-mail: commits-h...@druid.apache.org