From a218de2e4502c831f84961b87ee329f22ddbf14f Mon Sep 17 00:00:00 2001 From: Hodayah Zargari Date: Wed, 22 Jul 2026 14:13:16 +0300 Subject: [PATCH 1/7] Large message support - BigObjectIndexed32 --- data-explorer/kusto/management/alter-encoding-policy.md | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/data-explorer/kusto/management/alter-encoding-policy.md b/data-explorer/kusto/management/alter-encoding-policy.md index 990b248693..8ffafb75af 100644 --- a/data-explorer/kusto/management/alter-encoding-policy.md +++ b/data-explorer/kusto/management/alter-encoding-policy.md @@ -3,7 +3,7 @@ title: .alter column policy encoding command description: Learn how to use the `.alter column policy encoding` command to change the encoding policy. ms.reviewer: alexans ms.topic: reference -ms.date: 08/11/2024 +ms.date: 07/22/2026 --- # .alter column policy encoding command @@ -44,6 +44,7 @@ The following table contains the possible values for the *EncodingPolicyType* pa | `Identifier` | Suitable for columns that have data that represents ID-like information (for example, guids). This policy applies the required index for this column to gain both query performance and reduce size in the storage. | | `BigObject` | Suitable for columns of dynamic or string type, which holds large objects. For example, the output of [hll aggregate function](../query/hll-aggregation-function.md). This policy disables the index of this column and overrides `MaxValueSize` property in the encoding Policy to 2 MB. | | `BigObject32` | Similar to `BigObject` in terms of target scenarios. Overrides `MaxValueSize` property in the encoding Policy to 32 MB. | +| `BigObjectIndexed32` | Similar to `BigObject32` in terms of target scenarios. This policy enables indexing large, ingested data. | | `Vector16` | This profile is designed for storing vectors of floating-point numbers in 16 bits precision utilizing the [Bfloat16](https://en.wikipedia.org/wiki/Bfloat16_floating-point_format) instead of the default 64 bits. It is highly recommended for storing ML vector embeddings as it reduces storage requirements by a factor of 4 and accelerates vector processing functions such as [series_dot_product()](../query/series-dot-product-function.md) and [series_cosine_similarity()](../query/series-cosine-similarity-function.md), by orders of magnitude. | | `Null` | Sets the current default encoding policy to the column and clears the previous encoding policy profile. | From 85fc7a6345952fff7f6391add3644b6f4078ba91 Mon Sep 17 00:00:00 2001 From: Hodayah Zargari Date: Sun, 2 Aug 2026 13:11:59 +0300 Subject: [PATCH 2/7] guys comment --- data-explorer/kusto/management/alter-encoding-policy.md | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/data-explorer/kusto/management/alter-encoding-policy.md b/data-explorer/kusto/management/alter-encoding-policy.md index 8ffafb75af..c4fce53138 100644 --- a/data-explorer/kusto/management/alter-encoding-policy.md +++ b/data-explorer/kusto/management/alter-encoding-policy.md @@ -3,7 +3,7 @@ title: .alter column policy encoding command description: Learn how to use the `.alter column policy encoding` command to change the encoding policy. ms.reviewer: alexans ms.topic: reference -ms.date: 07/22/2026 +ms.date: 08/02/2026 --- # .alter column policy encoding command @@ -48,6 +48,9 @@ The following table contains the possible values for the *EncodingPolicyType* pa | `Vector16` | This profile is designed for storing vectors of floating-point numbers in 16 bits precision utilizing the [Bfloat16](https://en.wikipedia.org/wiki/Bfloat16_floating-point_format) instead of the default 64 bits. It is highly recommended for storing ML vector embeddings as it reduces storage requirements by a factor of 4 and accelerates vector processing functions such as [series_dot_product()](../query/series-dot-product-function.md) and [series_cosine_similarity()](../query/series-cosine-similarity-function.md), by orders of magnitude. | | `Null` | Sets the current default encoding policy to the column and clears the previous encoding policy profile. | +> [!NOTE] +> The `BigObjectIndexed32` encoding policy is relevant for columns of the [string](../query/scalar-data-types/string.md) or [dynamic](../query/scalar-data-types/dynamic.md) data type. Setting this policy raises the per-column maximum value size for that string or dynamic column to 32 MiB, up from the default 1 MiB. Values above the limit are still capped: string values are truncated to the limit, and dynamic values are replaced with `null`. Regardless of whether this policy is set, the maximum record (row) size remains 64 MiB. + ## Example ```kusto From 38b72fbcabee9336f80d1122b764fd1457e85a59 Mon Sep 17 00:00:00 2001 From: Hodayah Zargari Date: Thu, 6 Aug 2026 12:34:11 +0300 Subject: [PATCH 3/7] note modified --- data-explorer/kusto/management/alter-encoding-policy.md | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/data-explorer/kusto/management/alter-encoding-policy.md b/data-explorer/kusto/management/alter-encoding-policy.md index c4fce53138..0f9d897ba6 100644 --- a/data-explorer/kusto/management/alter-encoding-policy.md +++ b/data-explorer/kusto/management/alter-encoding-policy.md @@ -3,7 +3,7 @@ title: .alter column policy encoding command description: Learn how to use the `.alter column policy encoding` command to change the encoding policy. ms.reviewer: alexans ms.topic: reference -ms.date: 08/02/2026 +ms.date: 08/06/2026 --- # .alter column policy encoding command @@ -44,12 +44,13 @@ The following table contains the possible values for the *EncodingPolicyType* pa | `Identifier` | Suitable for columns that have data that represents ID-like information (for example, guids). This policy applies the required index for this column to gain both query performance and reduce size in the storage. | | `BigObject` | Suitable for columns of dynamic or string type, which holds large objects. For example, the output of [hll aggregate function](../query/hll-aggregation-function.md). This policy disables the index of this column and overrides `MaxValueSize` property in the encoding Policy to 2 MB. | | `BigObject32` | Similar to `BigObject` in terms of target scenarios. Overrides `MaxValueSize` property in the encoding Policy to 32 MB. | -| `BigObjectIndexed32` | Similar to `BigObject32` in terms of target scenarios. This policy enables indexing large, ingested data. | +| `BigObjectIndexed32` | Similar to `BigObject32` in terms of target scenarios. This policy enables the index of large data in this column. | | `Vector16` | This profile is designed for storing vectors of floating-point numbers in 16 bits precision utilizing the [Bfloat16](https://en.wikipedia.org/wiki/Bfloat16_floating-point_format) instead of the default 64 bits. It is highly recommended for storing ML vector embeddings as it reduces storage requirements by a factor of 4 and accelerates vector processing functions such as [series_dot_product()](../query/series-dot-product-function.md) and [series_cosine_similarity()](../query/series-cosine-similarity-function.md), by orders of magnitude. | | `Null` | Sets the current default encoding policy to the column and clears the previous encoding policy profile. | > [!NOTE] -> The `BigObjectIndexed32` encoding policy is relevant for columns of the [string](../query/scalar-data-types/string.md) or [dynamic](../query/scalar-data-types/dynamic.md) data type. Setting this policy raises the per-column maximum value size for that string or dynamic column to 32 MiB, up from the default 1 MiB. Values above the limit are still capped: string values are truncated to the limit, and dynamic values are replaced with `null`. Regardless of whether this policy is set, the maximum record (row) size remains 64 MiB. +> The `BigObject`, `BigObject32`, and `BigObjectIndexed32` encoding policies are relevant for columns with [string](../query/scalar-data-types/string.md) or [dynamic](../query/scalar-data-types/dynamic.md) data types. +> Setting the `BigObjectIndexed32` policy raises the per-column maximum value size for that string or dynamic column to 32 MB, up from the default 1 MB. Values above the limit are still capped: string values are truncated to the limit ( MaxValueSize property of the default policy is 1 MB), dynamic values are replaced with null. Regardless of whether this policy is set, the maximum record (row) size remains 64 MB. ## Example From 4a3ff6a605db3b28d4e0add5186cf78e53a56d7b Mon Sep 17 00:00:00 2001 From: Hodayah Zargari Date: Thu, 6 Aug 2026 14:08:17 +0300 Subject: [PATCH 4/7] row size --- data-explorer/kusto/query/schema-entities/tables.md | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/data-explorer/kusto/query/schema-entities/tables.md b/data-explorer/kusto/query/schema-entities/tables.md index 609242b5dc..c221e527a9 100644 --- a/data-explorer/kusto/query/schema-entities/tables.md +++ b/data-explorer/kusto/query/schema-entities/tables.md @@ -3,7 +3,7 @@ title: Tables description: This article describes Tables. ms.reviewer: orspodek ms.topic: reference -ms.date: 08/11/2024 +ms.date: 08/06/2026 --- # Tables @@ -15,6 +15,7 @@ Tables occupy the same namespace as [stored functions](stored-functions.md). If > [!NOTE] > +> * The maximum record (row) size is 64 MB. > * The maximum limit of tables per database is 10,000. > * Table names are case-sensitive. For more information, see [Identifier naming rules](entity-names.md#identifier-naming-rules). From 69973c77703a516028bf880866310fa598a9c311 Mon Sep 17 00:00:00 2001 From: Hodayah Zargari Date: Thu, 6 Aug 2026 14:38:26 +0300 Subject: [PATCH 5/7] truncation note --- data-explorer/kusto/management/alter-encoding-policy.md | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/data-explorer/kusto/management/alter-encoding-policy.md b/data-explorer/kusto/management/alter-encoding-policy.md index 0f9d897ba6..42ca59f45e 100644 --- a/data-explorer/kusto/management/alter-encoding-policy.md +++ b/data-explorer/kusto/management/alter-encoding-policy.md @@ -49,8 +49,10 @@ The following table contains the possible values for the *EncodingPolicyType* pa | `Null` | Sets the current default encoding policy to the column and clears the previous encoding policy profile. | > [!NOTE] -> The `BigObject`, `BigObject32`, and `BigObjectIndexed32` encoding policies are relevant for columns with [string](../query/scalar-data-types/string.md) or [dynamic](../query/scalar-data-types/dynamic.md) data types. -> Setting the `BigObjectIndexed32` policy raises the per-column maximum value size for that string or dynamic column to 32 MB, up from the default 1 MB. Values above the limit are still capped: string values are truncated to the limit ( MaxValueSize property of the default policy is 1 MB), dynamic values are replaced with null. Regardless of whether this policy is set, the maximum record (row) size remains 64 MB. +> +> * The `BigObject`, `BigObject32`, and `BigObjectIndexed32` encoding policies are relevant for columns with [string](../query/scalar-data-types/string.md) or [dynamic](../query/scalar-data-types/dynamic.md) data types. +> * Setting the `BigObjectIndexed32` policy raises the per-column maximum value size for that string or dynamic column to 32 MB, up from the default 1 MB. Values above the limit are still capped: `string` values are truncated to the limit (`MaxValueSize` property of the default policy is 1 MB), `dynamic` values are replaced with null. +> * Regardless of whether this policy is set, the maximum record (row) size remains 64 MB. ## Example From c1db802154615a38a765f98685563d561a3bf1c5 Mon Sep 17 00:00:00 2001 From: Hodayah Zargari Date: Wed, 19 Aug 2026 12:16:52 +0300 Subject: [PATCH 6/7] Added preview note --- data-explorer/kusto/management/alter-encoding-policy.md | 9 ++++++--- data-explorer/kusto/query/schema-entities/tables.md | 4 ++-- 2 files changed, 8 insertions(+), 5 deletions(-) diff --git a/data-explorer/kusto/management/alter-encoding-policy.md b/data-explorer/kusto/management/alter-encoding-policy.md index 42ca59f45e..1d5ac8dc4c 100644 --- a/data-explorer/kusto/management/alter-encoding-policy.md +++ b/data-explorer/kusto/management/alter-encoding-policy.md @@ -1,9 +1,9 @@ --- title: .alter column policy encoding command description: Learn how to use the `.alter column policy encoding` command to change the encoding policy. -ms.reviewer: alexans +ms.reviewer: guregini ms.topic: reference -ms.date: 08/06/2026 +ms.date: 08/19/2026 --- # .alter column policy encoding command @@ -44,7 +44,7 @@ The following table contains the possible values for the *EncodingPolicyType* pa | `Identifier` | Suitable for columns that have data that represents ID-like information (for example, guids). This policy applies the required index for this column to gain both query performance and reduce size in the storage. | | `BigObject` | Suitable for columns of dynamic or string type, which holds large objects. For example, the output of [hll aggregate function](../query/hll-aggregation-function.md). This policy disables the index of this column and overrides `MaxValueSize` property in the encoding Policy to 2 MB. | | `BigObject32` | Similar to `BigObject` in terms of target scenarios. Overrides `MaxValueSize` property in the encoding Policy to 32 MB. | -| `BigObjectIndexed32` | Similar to `BigObject32` in terms of target scenarios. This policy enables the index of large data in this column. | +| `BigObjectIndexed32` (preview) | Similar to `BigObject32` in terms of target scenarios. This policy enables the index of large data in this column. | | `Vector16` | This profile is designed for storing vectors of floating-point numbers in 16 bits precision utilizing the [Bfloat16](https://en.wikipedia.org/wiki/Bfloat16_floating-point_format) instead of the default 64 bits. It is highly recommended for storing ML vector embeddings as it reduces storage requirements by a factor of 4 and accelerates vector processing functions such as [series_dot_product()](../query/series-dot-product-function.md) and [series_cosine_similarity()](../query/series-cosine-similarity-function.md), by orders of magnitude. | | `Null` | Sets the current default encoding policy to the column and clears the previous encoding policy profile. | @@ -54,6 +54,9 @@ The following table contains the possible values for the *EncodingPolicyType* pa > * Setting the `BigObjectIndexed32` policy raises the per-column maximum value size for that string or dynamic column to 32 MB, up from the default 1 MB. Values above the limit are still capped: `string` values are truncated to the limit (`MaxValueSize` property of the default policy is 1 MB), `dynamic` values are replaced with null. > * Regardless of whether this policy is set, the maximum record (row) size remains 64 MB. + [!INCLUDE [feature-preview-note](../includes/feature-preview-note.md)] + + ## Example ```kusto diff --git a/data-explorer/kusto/query/schema-entities/tables.md b/data-explorer/kusto/query/schema-entities/tables.md index c221e527a9..2db79e2503 100644 --- a/data-explorer/kusto/query/schema-entities/tables.md +++ b/data-explorer/kusto/query/schema-entities/tables.md @@ -1,9 +1,9 @@ --- title: Tables description: This article describes Tables. -ms.reviewer: orspodek +ms.reviewer: guregini ms.topic: reference -ms.date: 08/06/2026 +ms.date: 08/19/2026 --- # Tables From 4af050ee4e50265ef2f05d78eddeea2ad0b894d5 Mon Sep 17 00:00:00 2001 From: Hodayah Zargari Date: Wed, 19 Aug 2026 12:24:44 +0300 Subject: [PATCH 7/7] preview note --- data-explorer/kusto/management/alter-encoding-policy.md | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/data-explorer/kusto/management/alter-encoding-policy.md b/data-explorer/kusto/management/alter-encoding-policy.md index 1d5ac8dc4c..0ca23d4716 100644 --- a/data-explorer/kusto/management/alter-encoding-policy.md +++ b/data-explorer/kusto/management/alter-encoding-policy.md @@ -50,12 +50,11 @@ The following table contains the possible values for the *EncodingPolicyType* pa > [!NOTE] > +> * The `BigObjectIndexed32` encoding policy is currently in preview. > * The `BigObject`, `BigObject32`, and `BigObjectIndexed32` encoding policies are relevant for columns with [string](../query/scalar-data-types/string.md) or [dynamic](../query/scalar-data-types/dynamic.md) data types. > * Setting the `BigObjectIndexed32` policy raises the per-column maximum value size for that string or dynamic column to 32 MB, up from the default 1 MB. Values above the limit are still capped: `string` values are truncated to the limit (`MaxValueSize` property of the default policy is 1 MB), `dynamic` values are replaced with null. > * Regardless of whether this policy is set, the maximum record (row) size remains 64 MB. - [!INCLUDE [feature-preview-note](../includes/feature-preview-note.md)] - ## Example