From 085d2cb4614f7c8fd9feb5041981fa64e0a63f98 Mon Sep 17 00:00:00 2001 From: henryshere Date: Thu, 26 Sep 2024 18:12:21 +0800 Subject: [PATCH] Update performance_faq.md --- site/en/faq/performance_faq.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/site/en/faq/performance_faq.md b/site/en/faq/performance_faq.md index ecb3c4486..2f61c5e72 100644 --- a/site/en/faq/performance_faq.md +++ b/site/en/faq/performance_faq.md @@ -22,7 +22,7 @@ The following charts are results from a test running on the sift50m dataset and #### Why do queries sometimes take longer on smaller datasets? -Query operations are conducted on segments. indexes reduce the amount of time it takes to query a segment. If a segment has not been indexed, Milvus resorts to brute-force search on the raw data—drastically increasing query time. +Query operations are conducted on segments. Indexes reduce the amount of time it takes to query a segment. If a segment has not been indexed, Milvus resorts to brute-force search on the raw data—drastically increasing query time. Therefore, it usually takes longer to query on a small dataset (collection) because it has not built index. This is because the sizes of its segments have not reached the index-building threshold set by `rootCoord.minSegmentSizeToEnableindex`. Call `create_index()` to force Milvus to index segments that have reached the threshold but not yet been automatically indexed, significantly improving query performance.