From 0ab13c555e1784693e5bb8c3c568e6f79186a11b Mon Sep 17 00:00:00 2001 From: davidmyriel Date: Tue, 13 Aug 2024 18:07:28 -0700 Subject: [PATCH] fix language --- .../content/documentation/concepts/indexing.md | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/qdrant-landing/content/documentation/concepts/indexing.md b/qdrant-landing/content/documentation/concepts/indexing.md index 39920875c..8e9f911c6 100644 --- a/qdrant-landing/content/documentation/concepts/indexing.md +++ b/qdrant-landing/content/documentation/concepts/indexing.md @@ -424,7 +424,7 @@ await client.CreatePayloadIndexAsync( *Available as of v1.11.0* -By default all payload-related structures are stored in memory. In this way vector index can quickly access payload values during search. +By default all payload-related structures are stored in memory. In this way, the vector index can quickly access payload values during search. As latency in this case is critical, it is recommended to keep hot payload indexes in memory. There are, however, cases when payload indexes are too large or rarely used. In those cases, it is possible to store payload indexes on disk. @@ -552,7 +552,7 @@ The list will be extended in future versions. *Available as of v1.11.0* -Many vector search use-cases require multi-tenancy support. In multi-tenant scenario the collection is expected to contain multiple subsets of data, each subset belonging to a different tenant. +Many vector search use-cases require multitenancy. In a multi-tenant scenario the collection is expected to contain multiple subsets of data, where each subset belongs to a different tenant. Qdrant supports efficient multi-tenant search by enabling [special configuration](../guides/multiple-partitions/) vector index, which disables global search and only builds sub-indexes for each tenant. @@ -560,8 +560,8 @@ Qdrant supports efficient multi-tenant search by enabling [special configuration In Qdrant, tenants are not necessarily non-overlapping. It is possible to have subsets of data that belong to multiple tenants. -But knowing that the collection contains multiple tenants unlocks more opportunities for optimization. -For enabling Qdrant to optimize storage even further, you can enable tenant index for payload fields. +However, knowing that the collection contains multiple tenants unlocks more opportunities for optimization. +To optimize storage in Qdrant further, you can enable tenant indexing for payload fields. This option will tell Qdrant which fields are used for tenant identification and will allow Qdrant to structure storage for faster search of tenant-specific data. One example of such optimization is localizing tenant-specific data closer on disk, which will reduce the number of disk reads during search. @@ -672,7 +672,7 @@ await client.CreatePayloadIndexAsync( ``` -Tenant optimization is supported for following types: +Tenant optimization is supported for the following datatypes: * `keyword` * `uuid`