– Using the Search Optimization Service — Snowflake Documentation

Page Meta Tags

viewport width=device-width, initial-scale=1
theme-color #ffffff
snowflake-release 4.36.2

Page Headers

0 HTTP/1.1 200 OK
Content-Type text/html
Content-Length 78796
Connection close
Date Wed, 04 Nov 2020 07:49:33 GMT
Last-Modified Tue, 03 Nov 2020 13:17:31 GMT
ETag “b301ad42ce109d547eac8b75516e5d25”
Server AmazonS3
X-Cache Hit from cloudfront
Via 1.1 (CloudFront)
X-Amz-Cf-Pop YTO50-C3
X-Amz-Cf-Id 3hvZls18WNx2GPsp81-6PqBlkXuelsYRQajlsSSAhpVlEKLIENCkRw==

Keyword Frequency

search 172
optimization 148
table 129
service 74
span 64
data 47
that 44
from 33
pre 32
can 32

Keyword Cloud

Using the Search Optimization Service mdash Snowflake Documentation DOCUMENTATION Community Resources Blog LANGUAGES English Deutsch Fran ais Getting Started Introduction to Tutorials Videos amp Other Release Notes Connecting Loading Data into Unloading from Web Interface Virtual Warehouses Databases Tables Views Understanding Table Structures Working with Temporary and Transient External How Does Work Considering Solutions for Optimizing Query Performance What Access Control Privileges Are Needed For Identifying That Benefit From Queries Current Limitations of Types Supported By Adding a With Optimized Modifying Cloning Schema or Database Replicating Sharing Managing Costs Estimating Viewing Reducing Removing Property Examples Overview Secure Materialized Design Considerations Storage Binary Date Time Semi-structured Travel Fail-safe Continuous Pipelines Replication Failover Failback Sample Sets Securely in Your Account Security Developing Applications General Reference SQL Command Function Appendices Next Previous Docs raquo Preview Feature Open Available all accounts Please note billing implications when using this preview feature as described section Enterprise Edition This requires To inquire about upgrading please contact Support The search optimization service can significantly improve performance point lookup queries In Topic Affects Joins aims selective on large tables A query returns only one small number distinct rows Use case examples include Business users who need fast response times critical dashboards highly filters scientists are exploring data volumes looking specific subsets user register more is table-level property applies columns supported types see list further below lookups relies persistent structure that serves an optimized access path maintenance runs background responsible creating maintaining When you add table creates populates needed perform process populating take time depending size does work not block any concurrent operations updated example by loading new sets through DML automatically updates reflect changes If run hasn t been yet might slower but will always return up-to-date results transparent You don create warehouse maintains However do be aware because there cost storage compute resources See several ways optimize Related techniques Clustering Creating materialized views clustered unclustered Each these has different advantages speed following long they clustering key Range searches Equality single which contain expressions speeds equality enabled view both range well some sort subset included also used order define keys same source conjunction flattening JSON variant shows three optimizations have costs Cost Compute View remove must privileges OWNERSHIP privilege ADD SEARCH OPTIMIZATION schema contains GRANT ON SCHEMA lt name gt TO ROLE role use just SELECT additional Because it detected if appropriate querying designed certain sections explain how identify benefit works best conditions true being queried at least GB smaller e g less than enough justify Either frequently other primary cluster typically tens seconds At accessed filter operation k- k values determine either span class pre APPROX COUNT DISTINCT get approximate select approx count column col actual c test approximation consider generally faster cheaper uses kinds predicates constant Predicates IN conjunctions i AND improved adhere above suppose where em condition x y separately few many disjunctions OR each predicate isolation decisive evaluated before determined top level WHERE b directly joins filtering prior join Both decision made independently As indirectly base All performed support Column concatenation Analytical Casts Although supports implicit explicit casts cast numeric implicitly varchar explicitly Unsupported string considers original after result active currently fixed-size Fixed-point numbers INTEGER NUMERIC DATE TIME TIMESTAMP VARCHAR BINARY Currently floating semi-structured listed future follow steps Switch Run command ALTER TABLE IF EXISTS alter information SHOW TABLES verify added much LIKE output Verify indicates Check value PROGRESS specifies percentage so far first benefits appear immediately starts increasingly catches up current state Before working wait until fully Note optimizer chooses particular Users cannot control Choose web UI plan node part effects becomes invalid altered dropped renamed adding invalidate moved again become then back drop paths Undropping reestablishes retention period removed needs recreate There no way undrop clone database cloned zero-copy its corresponding CREATE empty copied replicated replication providers share shared consumers improvements impacts space depends upon multiple factors including NDVs extreme unique required s Typically however approximately consumes Maintaining Resource consumption higher high churn change These roughly proportional amount ingested changed Deletes ensures efficient credit usage your account Billing calculated -second increments Serverless Credits Consumption per hour Once enable Tip recommends starting slowly closely monitoring estimate SYSTEM ESTIMATE COSTS function general those interface Features carefully choosing addition reduce batching DELETE store most recent day week month trim deleting old cases able daily rather hourly INSERT UPDATE MERGE Batching statements recluster entire dropping reclustering re-adding DROP code creation Start replace id int date null Add An clause compatible individually joined still delete update set Ask Contact Report Doc Issue Platform Cloud Architecture Pricing Marketplace Healthcare Life Sciences Marketing Analytics Retail Education Developers Library Webinars Legal Explore News Trending About Leadership Board Careers Privacy Notice Site Terms Concard Drive San Mateo CA United States -SNOWFLK Inc Rights Reserved