> ## Documentation Index > Fetch the complete documentation index at: https://cockroachlabs.mintlify.site/llms.txt > Use this file to discover all available pages before exploring further. # Vector Indexes export const InternalLink = ({version, path = "", children, ...props}) => { let detectedVersion = version || "stable"; if (typeof window !== 'undefined' && !version) { const match = window.location.pathname.match(/\/docs\/([^/]+)/); if (match) { detectedVersion = match[1]; } } const normalizedPath = path.startsWith("/") ? path.slice(1) : path; return {children} ; }; A *vector index* enables efficient approximate nearest neighbor (ANN) search on high-dimensional `VECTOR` columns. Use vector indexes to improve the performance of similarity searches over large datasets, such as embeddings generated by machine learning models. This page describes how to create and use vector indexes on CockroachDB. `VECTOR` functionality is compatible with the [`pgvector`](https://github.com/pgvector/pgvector) extension for PostgreSQL. ## How do vector indexes work? CockroachDB vector indexes organize vectors into a hierarchical structure of partitions using [k-means clustering](https://en.wikipedia.org/wiki/K-means_clustering). This partition structure groups similar vectors together and enables efficient, [tunable](#tune-vector-indexes) ANN searches. When a query uses a vector index, CockroachDB explores a subset of partitions based on their proximity to the query vector. It then retrieves and evaluates a candidate set of vectors using the [configured distance metric](#comparisons) and returns the top nearest results. ## Enable vector indexes To enable the use of vector indexes, set the `feature.vector_index.enabled` cluster setting: ```sql theme={"theme":{"light":"catppuccin-mocha","dark":"catppuccin-mocha"}} SET CLUSTER SETTING feature.vector_index.enabled = true; ``` To enable the creation of vector indexes on non-empty tables, also disable the `sql_safe_updates` session setting. This allows vector indexes to be backfilled on existing rows, during which **table writes are blocked** to ensure vector index consistency. This blocking behavior is a [known limitation](#known-limitations) that is currently being tracked. Adding a vector index to a non-empty table can temporarily disrupt workloads that perform continuous writes. ```sql theme={"theme":{"light":"catppuccin-mocha","dark":"catppuccin-mocha"}} SET sql_safe_updates = false; ``` ## Create vector indexes To create a vector index, use the `CREATE VECTOR INDEX` statement: ```sql theme={"theme":{"light":"catppuccin-mocha","dark":"catppuccin-mocha"}} CREATE VECTOR INDEX ON {table} (column}); ``` You can also specify a vector index during table creation. For example: ```sql theme={"theme":{"light":"catppuccin-mocha","dark":"catppuccin-mocha"}} CREATE TABLE items ( department_id INT, category_id INT, embedding VECTOR(1536), VECTOR INDEX (embedding) ); ``` ### Define prefix columns You can create a vector index with one or more *prefix columns* to pre-filter the search space. This is especially useful for tables containing millions of vectors or more. ```sql theme={"theme":{"light":"catppuccin-mocha","dark":"catppuccin-mocha"}} CREATE TABLE items ( department_id INT, category_id INT, embedding VECTOR(1536), VECTOR INDEX (department_id, category_id, embedding) ); ``` A vector index is only used if each prefix column is constrained to a specific value in the query. For example: ```sql theme={"theme":{"light":"catppuccin-mocha","dark":"catppuccin-mocha"}} WHERE department_id = 100 AND category_id = 200 ``` You can filter on multiple prefix values using `IN`: ```sql theme={"theme":{"light":"catppuccin-mocha","dark":"catppuccin-mocha"}} WHERE (department_id, category_id) IN ((100, 200), (300, 400)) ``` The following example will not use the vector index: ```sql theme={"theme":{"light":"catppuccin-mocha","dark":"catppuccin-mocha"}} WHERE department_id = 100 AND category_id >= 200 ``` For an example, refer to [Create and query a vector index](#create-and-query-a-vector-index). ### Specify an opclass When you create a vector index, you can specify an *operator class* (opclass) that tells the index which `VECTOR` [distance metric](#comparisons) to accelerate. The following opclasses are available: * `vector_l2_ops` (default): Accelerate queries that use the L2 distance operator `<->`. * `vector_cosine_ops`: Accelerate queries that use the cosine distance operator `<=>`. * `vector_ip_ops`: Accelerate queries that use the negative inner product operator `<#>`. If not specified, `vector_l2_ops` is used by default. To accelerate cosine or inner-product searches, specify the corresponding opclass when you create the vector index. For an example, to build a cosine-optimized index: ```sql theme={"theme":{"light":"catppuccin-mocha","dark":"catppuccin-mocha"}} CREATE TABLE items ( department_id INT, category_id INT, embedding VECTOR(1536), VECTOR INDEX embed_idx (embedding vector_cosine_ops) ); ``` ## Comparisons Vector indexes on `VECTOR` columns support the following comparison operators. Whether an operator is *accelerated* depends on the [opclass](#specify-an-opclass) that you specified upon index creation. * **L2 distance** (`<->`): Use when you want the true geometric distance, such as in spatial or physical models where absolute positioning matters. * **Cosine distance** (`<=>`): Use when you only care about directional similarity, like in semantic text matching or clustering. Ideal for [retrieval-augmented generation (RAG)](https://en.wikipedia.org/wiki/Retrieval-augmented_generation) use cases involving pretrained embedding models that either normalize vectors or are trained with a cosine similarity loss. * **Negative inner product** (`<#>`): Use when both the magnitude and direction of vectors matter, such as in scoring or preference modeling. Operators are used in the `ORDER BY` clause when ranking results by vector similarity. Refer to the [Example](#example). ## Tune vector indexes The following vector index parameters are tunable: * Search beam size, using `vector_search_beam_size` * Partition size, using `min_partition_size` and `max_partition_size` * Build beam size, using `build_beam_size` (not recommended) For guidelines on how these parameters affect search accuracy and write performance, refer to [Tuning considerations](#tuning-considerations). Set the following storage parameters when you create a vector index: * `min_partition_size`: Minimum number of vectors in a partition before it is merged with another partition. This value must be greater than or equal to `1`, up to a maximum of `1024`. The default value is `16`. * `max_partition_size`: Maximum number of vectors in a partition before it is split into smaller partitions. This must be at least 4 times the value of `min_partition_size`, up to a maximum of `4096`. The default value is `128`. * `build_beam_size`: Beam size for index build (how many branches of the k-means tree are explored when assigning a vector to a partition). The default value is `8`. Cockroach Labs does not recommend changing this setting. It is more effective to increase `vector_search_beam_size`. For example, the following statement creates a vector index with a custom partition size: ```sql theme={"theme":{"light":"catppuccin-mocha","dark":"catppuccin-mocha"}} CREATE VECTOR INDEX ON items (category, embedding) WITH (min_partition_size=16, max_partition_size=128); ``` Set the `vector_search_beam_size` session setting to determine how many vector partitions will be considered during query execution. The default value is `32`, which represents the number of partitions that are explored at each level of the k-means tree. For example: ```sql theme={"theme":{"light":"catppuccin-mocha","dark":"catppuccin-mocha"}} SET vector_search_beam_size = 16; ``` ### Tuning considerations Partition size and beam size interact to control both the precision of nearest neighbor search and the cost of maintaining the index. You can improve the accuracy of vector searches by increasing either the search beam size or partition size: * A larger search beam size improves accuracy by exploring more partitions, which increases the number of candidate vectors evaluated. * A larger partition size improves accuracy by placing more vectors in each partition, which increases the number of candidate vectors retrieved for a given beam size. It also improves write performance by reducing the frequency of partition splits and merges. In both cases, read latency and CPU usage increase with size. Changing `build_beam_size` offers little to no practical benefit compared to adjusting `vector_search_beam_size`. In most cases, tuning `build_beam_size` will not yield meaningful accuracy improvements and can negatively impact index construction performance. #### Search accuracy To explore more partitions during a search, increase `vector_search_beam_size`. This improves search accuracy by evaluating more of the index, but increases CPU usage and read latency because more partitions are scanned. To group more vectors into each partition, increase the partition size with `min_partition_size` and `max_partition_size`. This improves search accuracy by causing more candidate vectors to be evaluated per partition. However, it increases CPU usage and increases read latency because each partition contains more vectors. For the same reason, you can often reduce `vector_search_beam_size` without sacrificing accuracy. Search accuracy is highly dependent on workload factors such as partition size, the number of `VECTOR` dimensions, how well the embeddings reflect semantic similarity, and how vectors are distributed in the dataset. Experiment with `vector_search_beam_size`, `min_partition_size`, and `max_partition_size` on a representative dataset to find the optimal setting for your workload. You can improve search accuracy for filtered queries by creating a [vector index with a prefix column](#create-and-query-a-vector-index). #### Write performance Because a larger partition size leads to fewer partition splits and merges, it enables faster insert performance. To optimize writes, increase the partition size with `min_partition_size` and `max_partition_size`. A larger partition size leads to fewer partition splits and merges, resulting in faster insert performance. Large batch inserts of `VECTOR` types can cause performance degradation. When inserting vectors, batching should be avoided. For an example, refer to Create and query a vector index. ## Example In the following example, a vector index with a prefix column is used to optimize searches on a `VECTOR` column in 512 dimensions. The example table is populated with 156,541 example rows, using a Python script to quickly insert the vectors. ### Before you begin * Create a virtual `python3` environment and install `psycopg[binary]`: ```shell theme={"theme":{"light":"catppuccin-mocha","dark":"catppuccin-mocha"}} python3 -m venv ~/venv source ~/venv/bin/activate pip install psycopg[binary] ``` * Download the Python script and sample data: ```shell theme={"theme":{"light":"catppuccin-mocha","dark":"catppuccin-mocha"}} curl -O https://vector-examples.s3.us-east-2.amazonaws.com/fast_insert.py curl -O https://vector-examples.s3.us-east-2.amazonaws.com/clip_embeddings_with_customers.csv ``` Ensure that the `.py` and `.csv` files are located in the same directory. ### Create and query a vector index 1. Start a single-node cluster: ```shell theme={"theme":{"light":"catppuccin-mocha","dark":"catppuccin-mocha"}} cockroach start-single-node \ --insecure \ --listen-addr=localhost:26257 \ --http-addr=localhost:8080 ``` 2. In a separate terminal, open a SQL shell on the cluster: ```shell theme={"theme":{"light":"catppuccin-mocha","dark":"catppuccin-mocha"}} cockroach sql --insecure ``` 3. [Enable vector indexes](#enable-vector-indexes) on the cluster: ```sql theme={"theme":{"light":"catppuccin-mocha","dark":"catppuccin-mocha"}} SET CLUSTER SETTING feature.vector_index.enabled = true; ``` 4. Create an `items` table that includes a `VECTOR` column called `embedding`, along with a vector index that uses `customer_id` as the prefix column: ```sql theme={"theme":{"light":"catppuccin-mocha","dark":"catppuccin-mocha"}} CREATE TABLE items ( id UUID PRIMARY KEY DEFAULT gen_random_uuid(), customer_id INT NOT NULL, name TEXT, embedding VECTOR(512), VECTOR INDEX (customer_id, embedding) ); ``` 5. In another terminal, run the Python script to insert the `clip_embeddings_with_customers.csv` data into the `items` table: ```shell theme={"theme":{"light":"catppuccin-mocha","dark":"catppuccin-mocha"}} python fast_insert.py ``` This process should take approximately 5-10 minutes. 6. When the script is finished executing, verify that `items` is populated: ```sql theme={"theme":{"light":"catppuccin-mocha","dark":"catppuccin-mocha"}} SHOW TABLES; ``` ``` schema_name | table_name | type | owner | estimated_row_count | locality --------------+------------+-------+-------+---------------------+----------- public | items | table | root | 156541 | NULL ``` 7. Perform a vector search using the `<->` L2 distance [operator](#comparisons). Include the `WHERE` clause to query only the vectors associated with a given `customer_id`, thus narrowing the vector search space: ```sql theme={"theme":{"light":"catppuccin-mocha","dark":"catppuccin-mocha"}} SELECT id, name, embedding FROM items WHERE customer_id = 1 ORDER BY embedding <-> '[0.00644,-0.00866,-0.00977,0.02129,0.02191,-0.02144,-0.01439,-0.08154,-0.0587,0.06555,-0.00967,0.00842,0.02094,-0.02939,0.0192,-0.007866,-0.10443,0.0015745,-0.0217,-0.02982,-0.08563,-0.02203,0.0501,-0.003183,-0.01802,0.0655,0.01355,0.01678,-0.03091,-0.02046,0.01982,0.01646,0.01002,-0.0065,0.01979,0.004524,0.01073,-0.05038,-0.03114,-0.101,-0.04953,-0.02039,-0.023,-0.02094,-0.00508,0.0516,0.04684,0.04715,-0.0394,-0.007435,0.01808,-0.0278,0.05685,0.0329,0.00896,0.010826,-0.005516,-0.010315,-0.02145,0.02296,0.1049,-0.005733,0.03433,0.001492,0.02802,-0.0556,0.006145,0.1013,0.0001343,-0.03726,-0.006786,0.00227,-0.04358,0.04694,0.03943,0.03278,-0.0363,-0.01343,-0.03497,-0.0328,0.01595,0.05524,-0.01823,0.001669,-0.00695,0.02367,0.01013,0.002083,0.01569,-0.02184,0.02864,0.01169,-0.601,0.02724,0.02263,-0.02649,-0.0441,0.00481,-0.03882,0.03064,0.01927,0.03833,-0.010216,0.004444,0.00893,0.0193,0.03137,0.03958,0.01245,0.02023,0.02443,-0.0112,-0.0187,-0.03604,0.004948,-0.01305,-0.03836,0.02914,0.01202,-0.00994,0.0363,-0.02258,0.001203,-0.00885,-0.01624,-0.0581,0.0009136,-0.003847,-0.0367,0.02737,0.00599,-0.0483,-0.0118,0.0759,0.00767,0.003883,0.03516,0.02686,-0.03726,-0.00888,-0.01439,0.001924,-0.01883,0.002098,0.008156,0.02428,0.00918,0.0635,-0.0732,0.02007,-0.04968,-0.007645,0.0289,0.02925,0.0748,-0.01974,-0.005478,0.01422,-0.00888,0.00675,-0.0424,-0.0303,0.04932,0.02055,0.02792,0.0472,-0.003717,0.002607,-0.02156,-0.01054,0.01159,0.004093,0.0001816,-0.01353,-0.01317,-0.05783,-0.03,0.01709,-0.01575,0.0304,0.02939,-0.01793,0.00964,0.0189,-0.01825,0.01331,-0.00656,-0.04044,-0.02902,-0.01347,0.004368,0.001875,0.005116,0.01038,-0.00973,0.002888,0.000518,0.02673,0.0642,-0.03763,-0.02138,-0.0416,0.00846,-0.02809,0.0316,0.001358,-0.02815,-0.001282,-0.01662,0.02885,-0.05585,0.01015,-0.001081,0.03366,0.013054,-0.001288,-0.003696,0.0241,0.0399,-0.00688,-0.0214,0.05804,-0.02957,0.001987,-0.05627,0.02072,-0.01066,-0.03433,0.01738,0.009476,-0.04984,-0.01028,-0.01958,-0.01398,-0.002432,-0.007023,0.00878,-0.01415,-0.007385,0.0034,0.01254,0.01034,0.00281,-0.01055,-0.03906,0.1171,0.00905,0.04788,-0.0153,0.0007515,-0.02356,0.02296,-0.01756,0.02966,-0.011826,-0.005955,0.01316,0.02348,0.163,-0.01337,0.04013,0.02695,-0.067,-0.01836,-0.01407,-0.03992,-0.02078,-0.03387,-0.009224,-0.02885,-0.0403,0.01034,-0.01936,-0.05518,0.01933,-0.02551,-0.01865,0.011566,0.004776,-0.006454,-0.00337,0.0293,-0.04312,0.02434,-0.01068,0.0129,0.1411,-0.009705,0.00439,0.007706,-0.01176,0.002438,0.01431,0.000449,-0.0555,-0.01441,-0.10535,-0.02295,-0.01897,0.010284,-0.012115,0.001859,0.03467,0.03192,0.001146,-0.02293,-0.02399,-0.005974,-0.010124,-0.0008497,-0.000651,0.01846,0.0758,-0.0278,0.006382,0.03943,0.01388,0.05035,0.01521,0.06116,0.03442,-0.01788,-0.0569,0.006584,0.01228,0.01238,-0.03021,-0.01166,-0.006516,-0.008354,-0.02957,-0.02144,0.03128,0.03708,-0.01941,0.010666,-0.001733,-0.01843,0.046,-0.002312,0.00544,-0.03162,-0.02313,0.03387,0.02907,-0.00806,0.007584,-0.005272,0.001955,0.01826,0.08746,0.00287,0.06085,-0.029,-0.002197,-0.1055,-0.0318,0.189,0.02068,-0.02332,0.009674,-0.02429,-0.01633,-0.00488,-0.0756,0.01032,0.06146,-0.0911,-0.02176,0.03766,-0.01198,0.063,-0.014595,-0.01846,-0.00856,0.02599,0.0907,-0.01367,0.0222,-0.02231,-0.0008545,0.0555,-5.585e-05,0.005394,0.02129,0.011,0.02121,-0.01152,-0.00693,0.1929,-0.01228,0.02162,0.01978,0.01041,0.02802,-0.02838,0.04248,0.02315,0.1019,-0.01604,-0.00924,0.01413,-0.02165,-0.02916,0.010605,-0.04526,-0.0052,-0.06186,-0.00414,0.02492,0.0158,0.01962,-0.010445,0.006237,-0.0173,0.000725,-0.04904,0.00871,0.02087,0.007633,0.04895,-0.01091,0.03452,0.05344,-0.03525,0.0241,-0.03035,-0.0248,0.1195,-0.005257,-0.02145,-0.004898,0.06647,0.0414,-0.02187,0.00893,0.01549,0.002043,-0.01074,0.00422,0.01991,0.04913,0.04034,0.00208,0.01962,-0.02525,-0.03244,-0.0376,-0.03757,0.003544,0.01556,0.0317,-0.05896,0.01587,0.03357,-0.0199,0.02716,-0.03397,0.02982,0.05362,0.0505,-0.0003583,0.00599,0.01162,0.014046,-0.01706,-0.0169,0.0216,-0.006832,0.04095,-0.01227,0.00589,0.01863,-0.00428,0.00327,0.00678,-0.02963,-0.012665,0.05237,-0.02408,0.0238,-0.02515,-0.01116,0.00577,-0.00711,0.05258,0.003374,0.04153,0.01464,-0.01508,-0.03598,-0.01016,0.0322,-0.0585,0.005238,-0.02141,-0.001159,0.010124,-0.02986,-0.02534,-0.005997,0.004307,-0.02657,0.01692,-0.03638,-0.05765,-0.03622,0.003136,-0.0616,0.07184,-0.006992,-0.007393]' LIMIT 3; ``` ``` id | name | embedding ---------------------------------------+------+--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- 2866cda3-dec1-4db6-92d3-e0409a7b60dd | NULL | [-0.03644,0.00482,0.006996,0.0397,0.02031,0.003174,0.05167,-0.02255,-0.0743,0.0247,0.02051,0.00207,0.02278,0.0117,0.04782,-0.02684,-0.1234,0.00639,-0.03387,-0.0494,-0.02823,-0.02385,0.03036,-0.0217,-0.04425,0.02248,-0.0113,0.00833,0.01473,0.02982,-0.03186,0.032,0.008354,-0.00673,0.003376,0.00646,-0.02066,-0.05078,-0.004715,-0.07306,-0.01863,-0.01497,-0.01407,-0.002058,0.04614,0.03128,0.02272,0.00572,0.000955,0.0413,0.005558,0.02249,0.00388,0.001359,0.01732,-0.02692,0.0137,0.02469,-0.01049,0.03384,0.0442,0.00991,-0.004353,0.00511,0.01435,-0.0434,0.00959,-0.04965,-0.0172,0.013466,0.01852,0.03873,-0.0102,0.00707,0.002037,0.03177,-0.02678,-0.03214,-0.03102,-0.03455,-0.005142,-0.05164,0.03394,-0.03485,0.0359,0.02263,0.009865,-0.0018,-0.000672,0.01395,0.04688,0.01104,-0.607,-0.011375,-0.006874,0.02104,-0.01787,-0.007572,-0.004025,0.0017395,0.0186,0.015205,-0.03333,0.0474,0.039,-0.00796,-0.02977,0.05194,0.006454,0.01056,-0.02045,-0.00486,-0.03824,0.006428,0.001293,-0.0268,-0.07,0.01863,0.001716,-0.0899,0.05594,-0.07916,0.01593,0.02011,0.013504,-0.04227,-0.04062,0.03958,-0.01197,-0.01006,0.0192,-0.0353,-0.002686,0.08795,0.002575,0.001193,-0.02109,-0.02652,-0.03099,0.03323,0.02167,-0.03955,0.004143,0.03265,0.02802,0.003092,-0.01134,0.003077,-0.004684,-0.02434,-0.02345,0.01712,0.0695,0.00811,0.03577,-0.03026,0.004063,0.0642,0.02876,-0.01639,-0.01843,-0.0696,0.004375,0.015434,0.01572,-0.016,0.06903,0.01205,0.02037,-0.03323,0.006317,0.009705,-0.002304,-0.0001507,-0.01113,-0.002771,-0.04773,0.014145,0.05124,0.04495,-0.001354,-0.0288,0.0207,0.02036,0.01369,-0.00684,-0.01397,-0.01591,-0.01817,-0.01554,-0.004925,-0.01067,0.002968,0.0596,0.005974,0.02592,-0.02792,0.005314,-0.00833,-0.00875,0.01863,-0.02574,0.02136,0.01362,0.002321,0.03162,-0.04065,0.000871,0.01346,0.0535,-0.011566,0.07074,0.007618,0.04932,0.01361,-0.001603,0.0289,0.03278,0.01909,0.02127,-0.04688,0.1057,-0.03738,0.01984,-0.04022,0.0111,-0.00466,-0.01808,0.05374,-0.01866,-0.02579,-0.0648,-0.0267,0.0648,-0.003239,0.03595,-0.00921,-0.003769,-0.00759,-0.0316,0.02364,-0.011086,0.03192,-0.0008087,-0.002354,-0.01492,-0.008965,0.03827,0.0317,0.05157,0.0001527,0.01423,0.006557,0.011765,-0.03833,0.00889,-0.06104,0.02867,0.1672,0.014336,-0.0161,0.015434,0.013176,-0.005596,0.02539,-0.01558,-0.05176,-0.01062,0.009155,-0.02776,-0.004932,0.01782,-0.0332,-0.03375,0.00864,-0.004425,0.01204,-0.03041,-0.01678,0.01637,0.02888,0.02235,0.003016,-0.005768,-0.00348,0.0488,-0.02907,-0.04785,-0.00077,0.03354,-0.00788,0.03424,0.01717,0.01683,-0.03534,0.0598,-0.04453,-0.004646,0.001023,0.0466,-0.011185,0.00532,0.02937,-0.01613,0.02042,-0.02347,-0.02483,-0.001838,0.02417,-0.001924,-0.0326,-0.01213,0.0877,-0.01035,-0.02525,0.02266,0.05524,0.0865,-0.00938,0.02348,0.01572,-0.004414,-0.03546,-0.0697,0.001506,0.02872,-0.00553,-0.009735,0.014336,0.01701,-0.01287,-0.01846,0.004284,-0.0694,-0.00635,-0.01692,-0.000953,-0.007477,-0.0008273,0.00807,0.00921,-0.011894,0.03998,0.0329,-0.03455,0.00799,0.05063,-0.00874,0.0002373,0.0226,0.0742,0.02228,0.04156,-0.03574,-0.02008,-0.0889,-0.03833,-0.01566,-0.000461,0.0349,0.0313,-0.02667,-0.01878,0.00832,-0.1144,0.02379,0.02774,-0.1357,0.03107,0.03583,0.00709,-0.00172,0.01248,-0.0072,0.02652,-0.01532,0.0684,-0.0369,3.62e-05,0.013435,-0.04797,0.0583,-0.0008154,-0.003048,-0.02094,0.001323,0.0204,-0.002422,0.013306,0.266,0.03226,0.04645,-0.005733,-0.03748,0.0342,-0.00557,0.007454,-0.01736,-0.03677,0.02315,-0.01697,0.02109,0.0515,-0.01747,0.041,-0.00872,-0.003311,-0.02701,-0.0381,-0.00983,-0.03528,0.00798,-0.009315,0.01622,-0.00243,-0.03995,-0.004265,-0.02551,-0.00563,0.01226,-0.0163,-0.0433,0.0512,0.11676,-0.02383,-0.00274,0.0289,-0.04498,0.1505,-0.003675,0.02513,0.03104,0.1508,0.00901,-0.005783,-0.03717,-0.03342,0.02475,-0.003098,-0.002007,0.01438,0.03046,-0.007656,-0.01273,-0.00438,-0.02122,-0.001858,-0.03226,-0.01491,0.006985,0.010124,-0.03723,-0.02298,-0.0126,0.0179,0.004795,0.01403,-0.006718,-0.01395,0.00429,-0.04205,0.000871,0.02074,-0.001945,0.07294,-0.00494,0.004997,-0.01564,-0.01724,0.00482,-0.02258,-0.003819,0.002863,0.00866,0.02623,-0.01046,0.01287,-0.02518,-0.005424,-0.01311,0.0372,-0.03754,0.03026,0.0295,-0.01616,-0.007473,0.01662,0.01788,0.007236,-0.02623,-0.03452,0.03406,0.002787,-0.04785,-0.0353,0.03503,-0.04565,0.01727,0.0008183,0.01628,-0.08044,-0.01057,0.02364,0.01581,0.01917,0.01651,0.003077,0.05594,0.004616,0.007362,0.01497,-0.01765] 513f4590-6752-4b2f-bb82-67d72c70f422 | NULL | [-0.02757,0.03363,-0.002504,-0.03143,0.010445,-0.01143,0.03604,-0.01811,0.005474,0.01024,0.003687,0.03348,-0.01122,-0.01259,-0.02325,-0.02614,-0.04807,-0.0165,-0.05173,-0.01097,-0.1339,0.005417,0.0317,0.00859,-0.0477,0.004086,0.002243,0.007233,0.001098,0.002834,-0.01364,0.03265,0.001186,-0.01025,0.0677,-0.02202,0.02513,0.01933,0.01411,-0.006374,0.02174,-0.001868,-0.02908,-0.046,-0.007904,-0.01037,0.05484,0.02788,-0.04868,0.04016,0.01578,-0.02301,0.007183,-0.001128,-0.02968,-0.004166,0.0356,-0.01477,0.005505,0.03494,0.02481,-0.01596,-0.03235,-0.001357,-0.002623,-0.03552,-0.02594,0.04453,-0.0286,-0.0174,0.011215,0.01537,0.003532,0.0501,0.00502,-0.00243,-0.02594,-0.03433,-0.009895,-0.004185,-0.01077,-0.009315,-0.02747,0.06113,0.01248,0.06012,-0.01628,-0.01474,0.0099,-0.03476,0.0008,-0.01125,-0.662,-0.05505,-0.01677,0.03537,0.01775,-0.002335,-0.1058,-0.0397,0.03235,0.0475,-0.01093,-0.00847,0.08856,0.03558,0.007225,-0.02354,-0.002796,0.01747,-0.000744,-0.071,-0.01648,-0.0071,0.01656,-0.02428,-0.002491,0.01855,0.0413,0.01563,0.0225,-0.0977,0.01176,-0.02167,0.02638,-0.06323,0.00852,0.001434,0.0004199,-0.00258,0.005215,0.003527,-0.01863,0.08484,0.003096,-0.01843,0.01115,0.002579,-0.03516,0.01261,0.01038,-0.003784,-0.014465,0.04703,-0.0346,-0.004364,-0.01103,0.03534,-0.0427,-0.03128,0.00154,0.003956,0.0849,0.01046,0.01349,-0.0556,0.03403,0.01761,-0.00954,-0.03506,-0.03653,-0.05484,-0.006783,0.0165,-0.01657,0.007904,0.0392,0.01097,0.02182,-0.04132,0.03023,0.01999,-0.04117,-0.011284,0.00849,-0.004578,-0.0568,0.007545,0.10474,0.02103,0.01229,-0.01224,-0.000572,0.02428,0.006638,-0.02933,-0.02191,0.02292,0.01004,-0.02577,0.005447,-0.0259,0.0088,0.043,0.006,0.00954,0.00209,-0.011185,0.00913,-0.02692,0.004944,0.02112,0.0145,0.005737,0.0094,0.02824,-0.02846,0.006325,-0.0215,0.008026,0.00893,0.04208,0.006428,0.03897,0.01944,-0.006557,0.0367,0.01634,-0.03888,0.003195,0.02547,0.163,-0.01797,0.032,-0.0434,-0.03114,0.03226,0.00644,0.05112,0.014175,-0.02707,-0.01158,0.04343,-0.004475,0.01518,-0.0232,-0.0006795,-0.01715,-0.03503,0.02348,0.01886,0.01442,0.01683,-0.02095,-0.02638,-0.009926,0.01022,0.01305,-0.004284,0.05594,-0.02388,0.01686,-0.02864,0.03714,-0.0329,-0.02339,-0.05585,0.01872,0.03062,-0.03558,-0.00523,0.001399,-0.0071,-0.1349,0.002829,-0.001425,0.002012,0.01302,0.0006456,-0.0382,-0.01566,0.0003338,-0.003765,-0.02191,0.013954,-0.0102,-0.02231,-0.05237,0.012085,0.01254,0.02873,0.00871,-0.0205,0.006245,0.00924,0.01877,-0.03482,-0.02145,0.00811,0.03677,-0.005703,0.04358,0.001885,-0.002228,-0.05258,0.008804,-0.04688,0.01163,0.02103,-0.0003643,0.03784,0.00559,0.01167,0.04205,0.04227,0.01569,0.02556,0.005676,-0.0214,0.015594,-0.02084,0.04962,0.0846,0.01317,0.00768,0.003391,0.04996,0.03001,0.0334,0.04266,0.02325,-0.04535,0.0067,-0.01184,0.01546,0.001275,0.00804,0.01394,-0.02615,0.01249,-0.0357,-0.000846,-0.04144,-0.04398,-0.005173,0.01415,0.003267,-0.01496,0.03235,-0.03256,0.0088,0.004997,0.01761,0.03683,-0.01743,0.01863,0.01194,-0.004982,0.02257,0.0003695,0.1049,0.013176,0.0371,-0.0324,-0.0156,-0.10675,-0.06903,0.09406,-0.007797,0.0532,0.10223,0.01275,-0.03586,0.03604,-0.01872,0.01129,0.01526,-0.1514,0.01288,0.0208,-0.03937,0.03928,-0.02327,-0.007015,-0.001682,0.01135,0.05408,-0.011375,-0.01837,0.01842,-0.02496,0.0001932,-0.01177,-0.01588,-0.0254,0.01057,0.04688,-0.002123,-0.04083,0.1378,-0.00748,0.0416,0.02693,-0.02605,0.02408,-0.04498,0.03806,0.0009537,-0.02771,-0.09216,0.005604,0.0324,0.02536,-0.02414,0.00976,-0.002401,0.02286,-0.02615,0.02904,-0.000424,0.02724,0.01851,-0.001937,-0.001205,-0.02917,0.009995,-0.07245,0.00454,-0.005375,-0.0172,-0.010345,-0.0474,0.02202,0.071,-0.0132,0.001837,-0.0014715,0.02368,0.0693,0.04727,0.0005713,0.0287,0.09436,-0.00534,-0.006596,-0.04266,-0.01209,0.0471,0.02582,-0.03198,0.01307,-0.0127,-0.001337,-0.006454,-0.02988,0.02591,-0.002998,-0.03296,-0.03275,-0.02698,0.0742,-0.01837,-0.03986,0.01974,0.02505,-0.000777,0.015335,-0.01904,-0.013916,0.01323,0.001882,0.03152,-0.007282,0.03372,0.03308,0.005276,0.00905,-0.044,-0.06903,0.001275,-0.03032,0.02202,-0.01657,-0.02832,-0.005085,-0.0206,-0.01184,-0.01447,-0.01884,-0.01501,-0.01541,-0.02992,0.004375,0.03162,-0.02385,0.02718,-0.01422,0.04727,0.0358,0.02664,-0.03604,-0.006096,-0.006454,-0.04654,-0.02336,0.012924,-0.008446,0.01808,-0.03503,0.02748,-0.03174,-0.02052,0.014275,0.002783,-0.02283,0.01484,-0.02454,0.003788,-0.0421,0.02864,-0.03177,-0.001041] 0f6db280-2f5c-46e2-b205-1a9789274fdd | NULL | [-0.000977,0.03008,0.01264,-0.0191,0.058,-0.02066,0.013016,-0.0173,-0.04626,-0.002918,0.00225,0.04648,0.08905,0.01449,0.02306,-0.003393,0.00471,0.0274,-0.01582,-0.001243,-0.0779,0.01241,0.04013,-0.0328,-0.03958,0.04245,0.0472,-0.00982,-0.00791,-0.01234,0.02104,0.01468,0.001149,-0.0204,0.04776,0.013565,0.02327,-0.03024,-0.01694,0.0321,0.006046,-0.03445,-0.04147,-0.036,0.02162,-0.001715,0.04346,0.0189,0.002394,-0.002722,0.03888,0.0211,0.007782,-0.00607,0.02922,0.04376,0.0466,-0.0001429,-0.04004,-0.01465,-0.0274,0.0381,0.01874,-0.02205,0.001417,-0.02008,0.00832,0.1,0.02376,-0.01936,0.03363,0.01427,-0.02505,0.02441,0.01863,-0.003431,-0.06323,0.006477,-0.01357,-0.02254,-0.001032,0.03503,-0.02559,0.001091,-0.01418,0.00471,-0.002892,-0.005604,0.04144,0.006493,0.002033,0.001527,-0.5815,-0.0408,-0.01397,-0.0269,0.03784,-0.003103,0.00254,-0.01134,0.0172,0.0203,-0.03253,0.01131,0.03918,-0.0003843,0.02423,-0.03638,-0.004627,0.00826,-0.02834,0.00881,-0.0187,0.02817,-0.02803,-0.03262,-0.01036,-0.01593,-0.001248,-0.04688,0.02213,-0.0231,0.01172,0.04047,-0.02185,-0.03723,-0.03105,-0.01598,-0.01811,0.0363,0.05566,-0.03192,0.05182,0.08356,-0.00711,-0.001823,0.0008454,0.02731,-0.04236,0.0138,0.02322,-0.00859,-0.02583,0.03041,0.00799,-0.0394,-0.000641,0.003445,-0.03802,-0.03464,-0.01482,0.0001334,0.1339,0.0358,0.002865,-0.063,0.02773,0.03336,0.0234,0.02739,-0.0608,-0.05664,0.01071,0.001357,-0.01196,0.02057,0.005497,0.001594,-0.03125,-0.03665,0.08344,0.02264,0.02051,-0.00868,-0.01581,0.0344,-0.12396,0.01107,0.04828,0.03357,0.02113,-0.0349,-0.01306,-0.01755,0.001984,-0.003275,-0.003826,0.03635,-0.03099,-0.00834,0.009575,-0.01292,-0.01805,0.02097,-0.0591,0.01932,-0.01383,0.00875,-0.02512,-0.07135,0.01423,-0.04843,0.02608,0.01518,0.010994,-0.011024,-0.05884,0.02287,-0.02579,0.0718,-0.107,-0.01247,-0.01697,0.02719,-0.003199,0.00372,0.0431,0.00832,-0.0082,0.01275,-0.0127,0.1322,-0.02583,-0.0173,-0.04126,-0.03934,0.006298,0.02167,0.06726,-0.03516,0.0186,-0.0665,-0.05072,0.05667,0.02922,-0.0575,-0.0249,0.01446,-0.009605,0.04834,0.01805,0.01423,0.01926,-0.0001755,0.0232,0.01836,0.02002,0.03305,-0.00636,0.04385,0.02626,0.001762,-0.00881,0.02336,-0.02011,-0.05426,-0.0093,0.0816,0.05774,-0.01006,0.02441,0.01507,0.0001404,-0.1952,-0.01779,0.006603,0.02126,-0.0118,0.01395,-0.02205,-0.0316,0.01348,-0.041,-0.02565,-6.056e-05,-0.0371,0.00999,0.003849,-0.01478,0.02855,0.01816,0.02171,-0.05563,0.02286,-0.01959,0.02803,0.19,-0.02783,-0.00093,0.03513,-0.02428,0.02867,0.02393,0.013405,-0.03455,0.011696,-0.12396,-0.02718,-0.00989,0.04935,0.0084,-0.005417,0.04126,0.02371,-0.001195,-0.0379,0.001037,0.02473,-0.03035,-0.003489,-0.03854,0.0676,0.08325,-0.01851,-0.02739,0.02844,0.02875,0.03857,-0.02931,0.09625,-0.01741,-0.1842,-0.02097,-0.015076,-0.00682,-0.02135,0.01218,0.0367,0.03955,-0.00705,-0.0797,0.01304,0.01625,-0.10333,-0.010376,-0.00648,0.0401,-0.003351,-0.004044,-0.02838,0.000996,-0.00687,-0.01151,0.04446,0.0206,-0.03183,0.02011,-0.04276,0.02425,-0.01828,0.0662,0.02696,0.04562,-0.02335,0.01321,-0.05106,-0.02187,0.01258,-0.01883,0.009155,0.07544,0.01332,-0.00706,0.02206,-0.09143,-0.005,0.02393,-0.0724,-0.00705,0.04205,-0.0215,0.03262,-0.02193,-0.02773,-0.0324,0.0167,0.08716,-0.02177,0.04605,0.005444,-0.03143,0.04556,0.005333,-0.02884,0.0079,0.01171,0.004105,-0.04163,-0.0182,0.02641,-0.02374,-0.002714,-0.0346,0.00935,0.01842,0.01767,0.006275,-0.02153,-0.01002,-0.01343,-0.0115,0.01061,-0.04398,0.003918,-0.01535,-0.02837,0.001749,-0.02745,-0.05524,-0.002453,-0.03714,0.01064,-0.01642,0.02625,-0.02249,-0.02637,-0.0637,-0.00944,0.04187,0.007,-0.006554,0.000269,-0.04156,0.01999,-0.02803,0.01653,0.001995,0.01417,0.01898,-0.003187,0.01182,0.05963,0.06696,0.02614,-0.01329,-0.01974,-0.01865,0.02406,-0.04266,-0.012535,0.01552,0.02473,-0.0001897,-0.02084,0.0084,-0.02176,0.006947,-0.04846,-0.04532,-0.0499,-0.01336,0.01921,-0.02257,-0.05414,0.05032,0.000645,0.03763,-0.02065,-0.03738,-0.02173,0.02533,0.02586,-0.005615,0.0093,0.01039,0.025,-0.0235,-0.0345,-0.02386,-0.003807,0.01294,-0.03625,-0.001739,-0.003138,0.04724,0.03128,0.01614,-0.004505,0.00232,0.02686,0.009514,-0.04816,0.00592,0.00842,-0.03784,0.04224,0.00503,0.0304,0.01193,-0.011894,-0.01842,0.0158,0.02986,-0.01178,-0.001757,0.0504,0.01653,0.039,-0.005116,0.000661,-0.05786,-0.01895,0.02576,-0.001175,-0.02713,0.0174,-0.03708,-0.00614,-0.003826,0.079,0.00843,0.002083] (3 rows) Time: 14ms total (execution 13ms / network 1ms) ``` 8. Use `EXPLAIN` to show how the vector index pre-filtered the vector search space: ```sql theme={"theme":{"light":"catppuccin-mocha","dark":"catppuccin-mocha"}} EXPLAIN SELECT id, name, embedding FROM items WHERE customer_id = 1 ORDER BY embedding <-> '[0.00644,-0.00866,-0.00977,0.02129,0.02191,-0.02144,-0.01439,-0.08154,-0.0587,0.06555,-0.00967,0.00842,0.02094,-0.02939,0.0192,-0.007866,-0.10443,0.0015745,-0.0217,-0.02982,-0.08563,-0.02203,0.0501,-0.003183,-0.01802,0.0655,0.01355,0.01678,-0.03091,-0.02046,0.01982,0.01646,0.01002,-0.0065,0.01979,0.004524,0.01073,-0.05038,-0.03114,-0.101,-0.04953,-0.02039,-0.023,-0.02094,-0.00508,0.0516,0.04684,0.04715,-0.0394,-0.007435,0.01808,-0.0278,0.05685,0.0329,0.00896,0.010826,-0.005516,-0.010315,-0.02145,0.02296,0.1049,-0.005733,0.03433,0.001492,0.02802,-0.0556,0.006145,0.1013,0.0001343,-0.03726,-0.006786,0.00227,-0.04358,0.04694,0.03943,0.03278,-0.0363,-0.01343,-0.03497,-0.0328,0.01595,0.05524,-0.01823,0.001669,-0.00695,0.02367,0.01013,0.002083,0.01569,-0.02184,0.02864,0.01169,-0.601,0.02724,0.02263,-0.02649,-0.0441,0.00481,-0.03882,0.03064,0.01927,0.03833,-0.010216,0.004444,0.00893,0.0193,0.03137,0.03958,0.01245,0.02023,0.02443,-0.0112,-0.0187,-0.03604,0.004948,-0.01305,-0.03836,0.02914,0.01202,-0.00994,0.0363,-0.02258,0.001203,-0.00885,-0.01624,-0.0581,0.0009136,-0.003847,-0.0367,0.02737,0.00599,-0.0483,-0.0118,0.0759,0.00767,0.003883,0.03516,0.02686,-0.03726,-0.00888,-0.01439,0.001924,-0.01883,0.002098,0.008156,0.02428,0.00918,0.0635,-0.0732,0.02007,-0.04968,-0.007645,0.0289,0.02925,0.0748,-0.01974,-0.005478,0.01422,-0.00888,0.00675,-0.0424,-0.0303,0.04932,0.02055,0.02792,0.0472,-0.003717,0.002607,-0.02156,-0.01054,0.01159,0.004093,0.0001816,-0.01353,-0.01317,-0.05783,-0.03,0.01709,-0.01575,0.0304,0.02939,-0.01793,0.00964,0.0189,-0.01825,0.01331,-0.00656,-0.04044,-0.02902,-0.01347,0.004368,0.001875,0.005116,0.01038,-0.00973,0.002888,0.000518,0.02673,0.0642,-0.03763,-0.02138,-0.0416,0.00846,-0.02809,0.0316,0.001358,-0.02815,-0.001282,-0.01662,0.02885,-0.05585,0.01015,-0.001081,0.03366,0.013054,-0.001288,-0.003696,0.0241,0.0399,-0.00688,-0.0214,0.05804,-0.02957,0.001987,-0.05627,0.02072,-0.01066,-0.03433,0.01738,0.009476,-0.04984,-0.01028,-0.01958,-0.01398,-0.002432,-0.007023,0.00878,-0.01415,-0.007385,0.0034,0.01254,0.01034,0.00281,-0.01055,-0.03906,0.1171,0.00905,0.04788,-0.0153,0.0007515,-0.02356,0.02296,-0.01756,0.02966,-0.011826,-0.005955,0.01316,0.02348,0.163,-0.01337,0.04013,0.02695,-0.067,-0.01836,-0.01407,-0.03992,-0.02078,-0.03387,-0.009224,-0.02885,-0.0403,0.01034,-0.01936,-0.05518,0.01933,-0.02551,-0.01865,0.011566,0.004776,-0.006454,-0.00337,0.0293,-0.04312,0.02434,-0.01068,0.0129,0.1411,-0.009705,0.00439,0.007706,-0.01176,0.002438,0.01431,0.000449,-0.0555,-0.01441,-0.10535,-0.02295,-0.01897,0.010284,-0.012115,0.001859,0.03467,0.03192,0.001146,-0.02293,-0.02399,-0.005974,-0.010124,-0.0008497,-0.000651,0.01846,0.0758,-0.0278,0.006382,0.03943,0.01388,0.05035,0.01521,0.06116,0.03442,-0.01788,-0.0569,0.006584,0.01228,0.01238,-0.03021,-0.01166,-0.006516,-0.008354,-0.02957,-0.02144,0.03128,0.03708,-0.01941,0.010666,-0.001733,-0.01843,0.046,-0.002312,0.00544,-0.03162,-0.02313,0.03387,0.02907,-0.00806,0.007584,-0.005272,0.001955,0.01826,0.08746,0.00287,0.06085,-0.029,-0.002197,-0.1055,-0.0318,0.189,0.02068,-0.02332,0.009674,-0.02429,-0.01633,-0.00488,-0.0756,0.01032,0.06146,-0.0911,-0.02176,0.03766,-0.01198,0.063,-0.014595,-0.01846,-0.00856,0.02599,0.0907,-0.01367,0.0222,-0.02231,-0.0008545,0.0555,-5.585e-05,0.005394,0.02129,0.011,0.02121,-0.01152,-0.00693,0.1929,-0.01228,0.02162,0.01978,0.01041,0.02802,-0.02838,0.04248,0.02315,0.1019,-0.01604,-0.00924,0.01413,-0.02165,-0.02916,0.010605,-0.04526,-0.0052,-0.06186,-0.00414,0.02492,0.0158,0.01962,-0.010445,0.006237,-0.0173,0.000725,-0.04904,0.00871,0.02087,0.007633,0.04895,-0.01091,0.03452,0.05344,-0.03525,0.0241,-0.03035,-0.0248,0.1195,-0.005257,-0.02145,-0.004898,0.06647,0.0414,-0.02187,0.00893,0.01549,0.002043,-0.01074,0.00422,0.01991,0.04913,0.04034,0.00208,0.01962,-0.02525,-0.03244,-0.0376,-0.03757,0.003544,0.01556,0.0317,-0.05896,0.01587,0.03357,-0.0199,0.02716,-0.03397,0.02982,0.05362,0.0505,-0.0003583,0.00599,0.01162,0.014046,-0.01706,-0.0169,0.0216,-0.006832,0.04095,-0.01227,0.00589,0.01863,-0.00428,0.00327,0.00678,-0.02963,-0.012665,0.05237,-0.02408,0.0238,-0.02515,-0.01116,0.00577,-0.00711,0.05258,0.003374,0.04153,0.01464,-0.01508,-0.03598,-0.01016,0.0322,-0.0585,0.005238,-0.02141,-0.001159,0.010124,-0.02986,-0.02534,-0.005997,0.004307,-0.02657,0.01692,-0.03638,-0.05765,-0.03622,0.003136,-0.0616,0.07184,-0.006992,-0.007393]' LIMIT 3; ``` ``` info -------------------------------------------------------------- distribution: local vectorized: true • top-k │ estimated row count: 3 │ order: +column9 │ k: 3 │ └── • render │ └── • lookup join │ table: items@items_pkey │ equality: (id) = (id) │ equality cols are key │ └── • vector search table: items@items_customer_id_embedding_idx target count: 3 prefix spans: [/1 - /1] ``` In the preceding output, `prefix spans: [/1 - /1]` shows that the search is limited to the part of the index where `customer_id = 1`. ## Known limitations * Large batch inserts of `VECTOR` types can cause performance degradation. When inserting vectors, batching should be avoided. For an example, refer to Create and query a vector index. * `IMPORT INTO` is not supported on tables with vector indexes. You can import the vectors first and create the index after import is complete. * The distance functions `vector_l1_ops`, `bit_hamming_ops`, and `bit_jaccard_ops` are not implemented. * Index acceleration with filters is only supported if the filters match prefix columns. * Index recommendations are not provided for vector indexes. ## See also * `VECTOR` * `CREATE INDEX` * Indexes