Scribd Logo
Upload

Welcome to Scribd!

  • 85 views

    Snowflake Training Presentation v1

    Uploaded by

    Akanksha Bokare

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content

    You might also like

    Snowflake Training Presentation v1

    Uploaded by

    Akanksha Bokare
    85 views111 pages

    Original Title

    Snowflake Training Presentation v1 (3)

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Report this Document

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    Download as pdf
    85 views111 pages

    Snowflake Training Presentation v1

    Uploaded by

    Akanksha Bokare

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    Download as pdf
    of 111
    C7™. e SECTION 1 — SNOWFLAKE INTRODUCTION ¢ INTRODUCTION ‘© CREATING SNOWFLAKE ACCOUNT ¢ TESTING SNOWFLAKE ¢ SNOWTLAKE ARCHITECTURE ‘¢ UNDERSTANDING THE WEB UI COMPONENTS. \RS OF FXPERIENCF IN VARIOUS. IOUS TECHNOLOGIES RELATED ING, REPORTING, ANALYTICS KE FROM QUITE LONG TIME > h 3 a INTRO TO SNOWFLAKE + CLOUD DATA WAREHOUSE * IN HIGH DEMAND BECAUSE OF NICHE FEATURES * PAY WHAT YOU USE * MODERN ARCHITECTURE — SEPARATES COMPUTE = AND STORAGE COST a * REMOVES SETUP OF INFRASTRUCTURE — BOTH Ss Land aol t Male Ol Aa t ° * ELASTIC AND HIGHLY SCALABLE 3) ® se @ SECTION 2 — DEEP DIVE INTO SNOWFLAKE ¢ INTRODUCTION ¢ SNOWFLAKE ARCHITECTURE © UNDERSTANDING THE WEB UI COMPONENTS. ¢ VIRTUAL WARCIIOUSES: ‘Standard (Getaut) STANDARD VS ECONOMY POLICY Preventsiminimies queuing by favoring startea ational ‘Cstrs over conenng cette keeping running dusters ful loaded rater tan stating ‘addvonal luster, which may resut in queries teing queved land taking longerte emplte The fest ster starts immediately wen ether a quer is queued othe ‘system detects that there's one more query than he curenty-unning _sterscan execute Each successive cluster wats to sar 20 ‘Harted For example your ‘warehouse is conigued with 10 max ‘luster, can take att 200+ seconds to start al 10 casters ‘enough query lad to keep the _lster busy frat east 8 minutes ‘Afer2ta 3 consecutive succesful checks (performed att minute Intervals), which determine whether the nad onthe least oaded cluster coud be edstrbuted te other clusters witout spinning up he hacks (performed att minute Intervals), which determine whether the load onthe eat loaded cuter coud be redstrbuted 1 Me other ster again ~ S C7™. eo SECTION 3 — SNOWFLAKE PRICING © PRICING INTRODUCTION ‘© WHAT IS SNOWFLAKE CREDIT @ SNOWFLAKE EDITIONS + SERVERLESS FEATURCS ¢ STORAGE COST © CLOUD SERVICES AND DATA TRANSFER COST SNOWFLAKE PRICING aS Y VERY IMPORTANT ASPECT WHILE WORKING ON SNOWFLAKE ‘SEPARATES COMPUTE AND STORAGE COST CHARGES BASED ON CONSUMED SNOWFLAKE CREDITS VALUE OF SNOWFLAKE CREDITS IS BASED ON SNOWFLAKE EDITION ‘WE WILL LEARN: TYPES OF SNOWFLAKE EDITIONS AND THEIR CREDITS COST SERVERLESS FEATURES WHAT IS A SNOWFLAKE CREDIT? STORAGE COST(ON-DEMAND /PRE-PURCHASED) VIRTUAL WAREHOUSES /COMPUTE COST DATA TRANSFER COST CLOUD SERVICES COST PRICING EXAMPLES °*2] Co o S AZ. wa Yu * A Snowflake credit is a unit of measure. + Snowflake credits are used to pay for the consumption of resources on Snowflake. User Receives $400 worth of free usage upon creation of Snowflake + Itis consumed only when a customer free trial account Ie uting reeourcer, auch ae when @ virtual warehouse is running, the cloud services layer is performing work, or serverless features are used. BN WHAT IS A SNOWFLAKE CREDIT? °“ 24 Co @ + Complete Data Wacehoure * secre Data Sharing across eons / + Premier Suppor 26365 Data Maretlace acess SNOWFLAKE EDITIONS + Search OptinizationSenice + Dynamic Dot Mashing + cena Oats Tkenaton| + Exerpeise+ + Pcicomplance * ata encryaton everywhere 1 Secret Secure using custemer managed eye 1 AWS Prhatenk support = customer dedicated metadata sore XN S, WZ Oo SERVERLESS FEATURES SNOWFLAKE OFFERS THE FOLLOWING ADDITIONAL SERVERLESS FEATURES THAT USE SNOWFLAKE-MANAGED COMPUTE RESQURCES AND CONSUME SNOWFLAKE CREDITS WHEN THEY ARE USED. 1+ haem es mtaed Ue Sontag Mts pil stern state pope ute ad materiaed resources hid ave pw Y YU + ciety speed op poitbolun aes + Uses Stoneman compute Y a / A Ss ~“ CHOOSE STORAGE TYPE WISELY © Y , ACTUAL cost u ‘ASSUMPTIONS SONGUMETIONT eter TOTAL COST PER MONTH 100GB OF STORAGE ‘WOULD BE REGUIRED PER _ OnE 2060 929/18 0.1 *93= $23 SFLFCTED PRF-PURCHASED WAARIAs ‘STORAGE PLAN 10GB OF STORAGE (Boum ee necure er a MONTH 20GB $40/18 0.02 * 40 = $0.8 Oo SELECTED ON-DEMAND KAVITA STORAGE PLAN A Ss ~“ CHOOSE STORAGE TYPE WISELY © Y , ACTUAL cost u ‘ASSUMPTIONS SONGUMETIONT eter TOTAL COST PER MONTH 100GB OF STORAGE ‘WOULD BE REGUIRED PER _ OnE 90GB 929/18 0.1 *93= $23 SFLFCTED PRF-PURCHASED WAARIAs ‘STORAGE PLAN 10GB OF STORAGE (Boum ee necure er a MONTH 90GB $40/18 0.09 * 40 = $3.6 Oo SELECTED ON-DEMAND KAVITA STORAGE PLAN ZZ STORAGE COST(ON- DEMAND/PRE- PURCHASED) (ON-DEMAND + THE EASIEST AND MOST FLEXIBLE WAY TO PURCHASE THE SNOWLAKE SERVICE IS ON DEMAND + CUSTOMERS ARE CHARGED A FIXED RATE FOR THE SERVICES THAT ‘ARE CONSUMED AND ARE BILLED IN ARREARS EVERY MONTH = COMMON PRICE ACROSS REGIONS: $40 PER MONTH PRE-PURCHASED CAPACITY = SNOWFLAKE PROVIDES CUSTOMERS THE OPTION TO PRE-PURCHASE CAPACITY. A CAPACITY PURCHASE IS A SPECIFIC DOLLAR COMMITMENT TO SNOWFLAKE = COMMON PRICE ACROSS REGIONS: $23 PER MONTH {ARE NEW AND UNSURE ABOUT THIS, IS TO FIRST OPT FOR THE ON-DEMAND, AND THEN SWITCH TO PRE-PURCHASED. ONCE THE ON-DEMAND CYCLE STARTS, MONITOR THE RESOURCE USAGE FOR A MONTH OR TWO, AND ONCE YOU HAVE A GOOD IDEA FOR YOUR MONTHLY DATA WAREHOUSING REQUIREMENTS, SWITCH TO A PRE- PURCHASED PLAN TO OPTIMIZE THE RECURRING MONTHLY CHARGES. °24 CO @ Y / AZ. wa oO e + SNOWFLAKE SUPPORTS A WIDE RANGE OF VIRTUAL Y WAREHOUSE SIZES: X-SMALL, SMALL, MEDIUM, LARGE, X-LARGE, 2X-LARGE, 3X-LARGE, AND 4X-LARGE. © THE SIZE OF THE VIRTUAL WAREHOUSE DETERMINES HOW FAST ‘QUERIES WILL RUN. Vi RTUAL * WHEN A VIRTUAL WAREHOUSE |S NOT RUNNING (THAT IS, WHEN. WAREHOUSE / IT IS SET TO SLEEP MODE}, IT DOES NOT CONSUME ANY COMPUTE SNOWFLAKE CREDITS. + THE DIFFERENT SIZES OF VIRTUAL WAREHOUSES CONSUME COST SNOWFLAKE CREDITS AT THE FOLLOWING RATES, BILLED BY THE SECOND WITH A ONE MINUTE MINIMUM. consumed 1 16 92s 128 perhour ) 706 i DATA TRANSFER COST wa + CUSTOMERS WHO WISH TO MOVE OR COPY THEIR DATA. BETWEEN REGIONS OR CLOUDS WILL INCUR DATA TRANSFER CHARGES. ‘+ FEATURES SUCH AS EXTERNAL TABLES, EXTERNAL FUNCTIONS AND DATA LAKE EXPORT MAY INCUR DATA TRANSFER CHARGES. + THE PRICING FOR THOSE CHARGES IS AS FOLLOWS FOR CUSTOMERS ON AMAZON WEB SERVICES OR MICROSOFT AZURE: “OQ e lasunedeeaeanaselll @ * CLOUD SERVICES RESOURCES ARE AUTOMATICALLY ASSIGNED BY SNOWFLAKE BASED ON THE REQUIREMENTS OF THE WORKLOAD. SERVICES + TYPICAL UTILIZATION OF CLOUD SERVICES (UP TO 10% OF COST DAILY COMPUTE CREDITS) IS INCLUDED FOR FREE * CUSTOMERS WILL NOT SEE INCREMENTAL CHARGES FOR CLOUD SERVICES USAGE. “OQ e ZZ OPTIMIZE/REDUCE SNOWFLAKE COSTS wa DEPENDING ON YOUR LOCATION, If IS IMPORTANT TO CHOOSE THE CLOUD REGION (UKE US BAST, US WEST, EIC. DEFENDING ON THE CLOUD FROVIDER) ‘WISELY, TO MINIMIZE LATENCY, TO HAVE ACCESS TO THE REQUIRED SET OF FEATURES, ETC. IF YOU ARE TO MOVE YOUR DATA TO A DIFFERENT REGION LATER, THERE ARE DATA TRANSFER COSTS ASSOCIATED WITH IT AT A PER TERABYTE SCALE, SO THE LARGER YOUR DATA STORE, THE MORE THE COSTS: IT CAN MAKE QUITE A DIFFERENCE TO THE COSTS INCURRED BY OPTIMALLY MANAGING THE OPERATIONAL STATUS OF YOUR COMPUTE CLUSTERS. THE FEATURES SUCH AS ‘AUTO SUSPENSION’ AND ‘AUTO RESUME’ SHOULD BE [MADE USE OF UNLESS THERE IS A BETTER STRATEGY TO ADDRESS THIS. THE WORKLOAD/DATA USAGE MONITORING AT AN ACCOUNT LEVEL, WAREHOUSE LEVEL, DATABASE OR TABLE LEVEL IS NECESSARY TO MAKE SURE THERE AREN'T UNNECESSARY QUERY OPERATIONS OR DATA STORAGE ‘CONTRIBUTING TO THE OVERALL MONTHLY COSTS. [MAKE SURE TO HAVE THE DATA COMPRESSED BEFORE STORAGE AS MUCH AS. POSSIBLE. THERE ARE INSTANCES, SUCH AS STORING DATABASE TABLES, WHERE SNOWFLAKE AUTOMATICALLY DOES DATA COMPRESSION, HOWEVER THIS IS NOT ALWAYS THE CASE, SO THIS IS SOMETHING TO BE MINDFUL OF AND TO BE MONITORED REGULARLY. ‘SNOWFLAKE WORKS BETTER WITH DATE OR TIMESTAMP COLUMNS STORED [AS SUCH RATHER THAN THEM BEING STORED AS TYPE VARCHAR, TRY TO MAKE MORE USE OF TRANSIENT TABLES AS THEY ARE NOT MAINTAINED INTHE HISTORY TABLES WHICH IN TURN REDUCES THE DATA STORAGE COSTS Re ane oe! | @ \ Qo Y ) QUIZ TIME - PRICING Q1. TOTAL COST IN DOLLARS/USD FOR 2TB OF DATA? Q2. TOTAL COST IN DOLLARS/USD FOR 13TB OF DATA IN AWS OHIO REGION? @3. TOTAL CREDITS CONSUMED AND CORRESPONDING COST FOR RUNNING VIRTUAL WARCIIOUSC (MEDIUM [4CREDITS/1 [OUR] - USED TOR CTL LOADS) RUNNING TOR 2HOURS PER DAY FOR 31DAYS IN A MONTH, IN AWS OHIO REGION? Q4. TOTAL CREDITS CONSUMED AND CORRESPONDING COST FOR RUNNING VIRTUAL WAREHOUSE (SMALL [2CREDITS/HOUR] USED BY USERS) RUNNING FOR 2HOURS PER DAY FOR 31DAYS IN A MONTH, IN AWS OHIO REGION? °“ 24 Co @ Y SS SECTION 4 — RESOURCE MONITORS ‘© RESOURCE MONITORS INTRODUCTION ‘¢ PROPERTIES OF RESOURCE MONITORS ‘¢ ACTIONS OR TRIGGERS OF RESOURCE MONITORS ¢ SUSPENSION OR RESUMPTION OF RESOURCE MONITORS ‘© CREATE RESOURCE MONITORS USING WEBUI CREATE RESOURCE MONITORS USING SQL S ZH oO e RESOURCE MONITORS + TO CONTROL COSTS AND AVOID UNEXPECTED CREDIT USAGE CAUSED BY RUNNING WAREHOUSES + CAN BE USED TO IMPOSE LIMITS ON THE NUMBER OF CREDITS THAT ARE CONSUMED BY VIRTUAL WAREHOUSES ‘+ WHEN LIMITS ARE REACHED AND/OR ARE APPROACHING, THE RESOURCE MONITOR CAN TRIGGER VARIGUS ACTIONS, SUCH AS SENDING ALERT NOTIFICATIONS AND /OR SUSPENDING THE WAREHOUSES. + CAN ONLY BE CEAIED BY ACLUUNI ADMINISIKAIOKS ow wf PROPERTIES OF RESOURCE MONITORS MONITOR LEVEL SCHEDULE (CREDIT QUOTA + rors me nunter oF suowras “EDT The NONTON FORTE SPECPED ‘EDU TEAL -AT THE SPEOHED NTERVAL, TS NUMBER ESET 280 MONTOR DOESNT NONITOR ANT CDT USAGE wy 0 W / RESOURCE MONITOR ACTIONS/TRIGGERS + ACTION SPECIFIES A THRESHOLD, AS A PERCENTAGE OF THE CREDIT QUOTA FOR THE RESOURCE MONITOR ‘+ ACTION TO PERFORM WHEN THE THRESHOLD IS REACHED WITHIN THE SPECIFIED INTERVAL ‘+ A RESOURCE MONITOR MUST HAVE AT LEAST ONE ACTION DEFINED; IF NO ACTIONS HAVE BEEN DEFINED, NOTHING HAPPENS [WHEN THE USED CREDITS REACH THE THRESHOLD NOTIFY & SUSPEND ‘SEND A NOTIFICATION (TO ALL ACCOUNT ADMINISTRATORS WITH NOTIFICATIONS. ENABLED] AND SUSPEND ALL ASSIGNED WAREHOUSES AFTER ALL STATEMENTS BEING EXECUTED BY THE WAREHOUSE(S) HAVE COMPLETED NOTIFY & SUSPEND IMMEDIATELY ‘SEND A NOTIFICATION (TO ALL ACCOUNT [ADMINISTRATORS WITH NOTIFICATIONS. ENABLED] AND SUSPEND ALL ASSIGNED WAREHOUSES IMMEDIATELY, WHICH CANCELS ANY STATEMENTS BEING EXECUTED BY THE WAREHOUSES AT THE TIME ~VA NOTIFY PERFORM NO ACTION, BUT SEND AN ALERT NOTIFICATION (TO ALL ACCOUNT ADMINISTRATORS WITH NOTIFICATIONS ENABLED) ~S XN Q EXAMPLE SALES MARKETING TECH FINANCE HR ‘An account-level resource monitor does not override resource monitor assignment for individual warehouses. If either the account resource moniter or the warehouse resource monitor reaches its defined threshold and @ suspend action has been defined, the warehouse Is suspended. Y ak SUSPEND OR =O WAREHOUSE ‘+ THE USED CREDITS FOR A RESOURCE MONITOR REFLECTS THE SUM OF ALL CREDITS CONSUMED BY ALL ASSIGNED WAREHOUSES WITHIN THE ‘SPECIFIED INTERVAL + IPA MONIIUK HAS A SUSPEND OX SUSFEND IMMEDIATELT ACIION DEFINED AND ITS USED CREDITS REACH THE THRESHOLD FOR THE ACTION, ANY WAREHOUSES ASSIGNED TO THE MONITOR ARE SUSPENDED AND (CANNOT BE RESUMED UNTIL ONE OF THE FOLLOWING CONDITIONS IS. “h THE NEXT INTERVAL, IF ANY, STARTS, AS DICTATED BY THE START DATE FOR THE MONITOR, 4h THE CREDIT QUOTA FOR THE MONITOR IS INCREASED. “P THE CREDIT THRESHOLD FOR THE SUSPEND ACTION IS INCREASED. ‘h THE WAREHOUSES ARE NO LONGER ASSIGNED TO THE MONITOR. ‘¢P THE MONITOR IS DROPPED. NOTE: WHEN CREDIT QUOTA THRESHOLDS ARE REACHED FOR A RESOURCE MONITOR, THE ASSIGNED WAREHOUSES MAY TAKE SOME TIME TO SUSPEND, EVEN WHEN THE ACTION IS SUSPEND IMMEDIATE, THEREBY CONSUMING ADDITIONAL CREDITS U SECTION 5 — MICRO PARTITIONING PARTITIONING IN TRADITIONA. WAREHOUSES. ‘© PARTITIONING IN SNOWFLAKE— MICRO PARTITIONING ‘¢ ADVANTAGES OF MICRO-PARTITIONING ‘¢ STRUCTURE OF MICRO-PARTITION ‘© QUERY PROCESSING IN SNOWFLAKE SS PARTITIONING IN TRADITIONAL DATA WAREHOUSES wa PARTITION IS A UNIT OF MANAGEMENT THAT IS MANIPULATED INDEPENDENTLY STATIC PARTITIONING OF LARGE TABLES HELPS TO ACHIEVE ACCEPTABLE PERFORMANCE, ENABLES BETTER SCALING. LIMITATIONS: ‘+ MAINTENANCE OVERHEAD + DATA SKEWNESS ‘+ DISPROPORTIONATELY-SIZED PARTITIONS °"“¥ QO € J % PARTITIONING IN SNOWFLAKE ro UNIQUE WAY OF PARTITIONING, KNOWN AS MICRO-PARTITIONING HAVE ALL BENEFITS, OVERCOMES LIMITATIONS OF STATIC PARTITIONING ‘MICRO-PARTITION CONTAINS BETWEEN 50 MB AND 500 MB OF UNCOMPRESSED. DATA EACH MICRO PARTITION CONTAINS GROUP OF ROWS IN A COLUMNAR STORAGE ‘SNOWFLAKE STORES METADATA ABOLT ALL ROWS STORED IN A MICRO-PARTITION, (0 THE RANGE OF VALUES FOR EACH OF THE COLUMNS IN THE MICRO-PARTITION (9 THE NUMBER OF DISTINCT VALUES. oe 7Q BENEFITS OF MICRO PARTITIONING iq SS ) Logical Structure Physical Structure y rd cd wa [| 92 oO Logical Structure Physical Structure y YZ. wa J UY DATA FILES STORED IN COLUMNAR STORAGE Y Size of file will be between 50 MB and 500 MB of uncompressed data WZ “Y © hf A aS J UY QUERY PROCESSING ON SNOWFLAKE (MICRO PARTITIONS /TABLE PLES Query sent to loud services layer SELECT TYPE, NAME, COUNTRY FROM EMPLOYEE WHERE DATE = "11/2" SELECT TYPE, NAME, COUNTRY FROWN EMPLOYEE WHERE DATE = "11/2" VIR ee ‘MICRO-PARTITIONS SELECTED. wa | a PRUNED WZ aS _J UY MICRO-PARTITIONS SELECTION AND PRUNING Y WZ aS J oO COLUMNS SELECTION © SELECT TYPE, NAME, COUNTRY FROM EMPLOYEE WHERE DATE = "11/2" SNOWFLAKE LOADS DATA INTO. THE MICRO-PARTITIONS AS AND WIIEN DATA IS INSERTED INTO THE TABLE ZF. aS Oo Y © SINGLE MICRO-PARTITION CAN HOLD ALL THE DATA ONLY IF MICRO-PARTITION SIZE IS BETWEEN 50-SOOMB SINGLE MICRO PARTITION Q SECTION 6 — CLUSTERING © CLUSTERING INTRODUCTION © CLUSTERING KEYS © CLUSTERING METADATA * CLUSTERING DEPTIL ‘@ AUTOMATIC RE-CLUSTERING DATA CLUSTERING APRULESD IU UPIIMIZE DAIA KEIKIEVAL DATA STORED IN TABLES IS SORTED/CRDERED ALONG NATURAL DIMENSIONS, CLUSTERING IS A KEY FACTOR IN QUERIES TO SORT/ORDER THE DATA FOR FASTER DATA RETRIEVAL TABLE DATA THAT IS NOT SORTED OR IS ONLY PARTIALLY SORTED MAY IMPACT QUERY PERFORMANCE, PARTICULARLY ON VERY LARGE TABLES DATA IS INSERTED/LOADED INTO A TABLE AND CLUSTERING METADATA IS, COLLECTED AND RECORDED FOR EACH MICRO-PARTITION. USING CLUSTERING INFORMATION, SNOWFIAKF AVOIDS UNNECESSARY SCANNING OF MICRO-PARTITIONS DURING QUERYING WHICH SIGNIFICANTLY ACCELERATES THE PERFORMANCE OF QUERIES A Y YS XN + SUBSET OF COLUMNS IN A TABLE (OR EXPRESSIONS ON A TABLE) ‘EXPLICITLY DESIGNATED TO CO-LOCATE THE DATA IN THE TABLEIN THE + USEFUL FOR VERY LARGE TABLES CLUSTERING © NO ADDITONAL ADMINISTRATION Is REQUIRED KEYS (©. FUTURE MAINTENANCE PERFORMED AUTOMATICALLY BY SNOWFLAKE + WHEN TO DEFINE CLUSTERING KEY: (©. QUERIES ON THE TABLE ARE RUNNING SLOWER THAN EXPECTED OR HAVE: NOTICEABLY DEGRADED OVER TIME, (0. THE CLUSTERING DEPTH FOR THE TABLE IS LARGE note clusmmne caw ar couPvranonaur poeie som cusmewien — (_7) oy QO © ) is ‘me ‘renee . eer) [cea « snontsoasnss acorn RAT ENA ATO RSET ay SI aaa cae “+ THE TOTAL NUMBER OF MICRO-PARTITIONS THAT COMPRISE THE TABLE + DATAIS SORTED AND STORED BY COLUMN 4 THE NUMBER OF MICRO-PARTITIONS CONTAINING VAILIES THAT ‘OVERLAP WITH EAC OTHER. Te ME HIRO ASIIONS IMAL AERO MED HO UE ‘+ THE DEPTH OF THE OVERLAPPING MICRO-PARITTIONS, 2. PRUNE BY COLLMN WITHIN THE REMAINING MCRC-PARTIIONS. ZF A ‘+ MEASURES THE AVERAGE DEPTH (1 OR GREATER) OF THE OVERLAPPING (MICRO-PARTITIONS FOR SPECIFIED COLUMNS IN A TABLE Y + SMALLER THE AVERAGE DEPTH, THE BETTER CLUSTERED THE TABLE + ADVANTAGES OF USING CLUSTERING DEPTH: ‘© MONITORING THE CLUSTERING “HEALTH” OF A LARGE TABLE, PARTICULARLY ‘OVER TIME AS DML IS PERFORMED ON THE TABLE. CLUSTERING (© DETERMINING WHETHER A LARGE TABLE WOULD BENEFIT FROM EXPLICITLY BEET DEFINING & CLUSTERING KEY. TADLE WITIINO MICRO-PARTITIONS IIAS A CLUSTERING DEFTII OF ZERO ‘+ CLUSTERING DEPTH IS NOT A PERFECT MEASURE TO CHECK IF TABLE IS WELL- (CLUSTERED. RATHER, CHECK QUERIES PERFORMANCE: (© IF QUERIES ON A TABLE ARE PERFORMING AS NEEDED OR EXPECTED, THE TABLE IS LIKELY WELL-CLUSTERED. oO ‘© IF QUERY PERFORMANCE DEGRADES OVER TIME, THE TABLES LIKELY NO LONGER WELL-CLUSTERED AND MAY BENEFIT FROM CLUSTERING. J °eYO e ) fo oo po oa.) f = «= as »J 1 fave oom | Micropparttions (Total) = § t 1 1 ‘AS DML OPERATIONS (INSERT, UPDATE, DELETE, MERGE, COPY) APE PERFORMED ON A CLUSTERED TABLE, THE DATA IN THE TABLE MIGHT BECOME LESS ‘CLUSTERED. PERIODIC/REGULAR RECLUSTERING OF THE TABLE IS REQUIRED TO MAINTAIN OPTIMAL CLUSTERING. RE-CLUSTERING & DURING RECLUSTERING, SNOWFLAKE USES THE CLUSTERING KEY FOR A CLUSTERED TABLE TO REORGANIZE THE COLUMN DATA, SO THAT RELATED RECORDS ARE RELOCATED TO THE SAME MICRO-PARTITION. erent Sa] a Go ZZ Yu SECTION 7 — QUERY HISTORY AND CACHING © QUERY HISTORY SQL QUERIES TO FETCH QUERY HISTORY © CACHING INTRODUCTION ¢ CACIING ~ DETAILED EXAMPLES ‘© TYPES OF CACHES wa 'e AVAILABLE FROM THE RIBBON ITEM — HISTORY ALLOWS VIEWING OF 14 DAYS OF QUERY HISTORY QUERY HISTORY CACHING 3 TYPES OF CACHES EXISTS ON SNOWFLAKE or Net = UPN ERIE Ola RSCTA SPEEDING UP RESULTS HISTORY AREA OF WEBUI HELPS TO VIEW USE OF US Ser NS Soe el wea eee eas eRe os ROSE 0 1 ols a RESULT CACHE TO PRESENT IT TO ANOTHER USER SELECT COUNT(*) FROM "SNOWFLAKE_SAMPLE_D ATA"."TPCDS_SFIOOTCL". “CALL_CENTER™ Metadata Results Warehouse Long-term Cache Cache SSD Cache Centralized Storage “Local” “Remote” € SELECT * FROM "SINOWFLAKE_SAMPLE_D ATA"."TPCDS_SFIOOTCL", "CALL_CENTER™; Results Warehouse Cache Cache SSD Cache “Local” J Long-term Centralized Storage “Remote” C SELECT * FROM "SINOWFLAKE_SAMPLE_D ATA"."TPCDS_SFIOOTCL", "CALL_CENTER™; Metadata Cache c RESULTS FS baa Results Warehouse Cache SSD Cache “Local” Long-term Centralized Storage “Remote” C o Y G . } _PATINE i CH i SELECT ! Geta) 1 b= | FROM i = oa "SNOWFLAKE_SAMPLE_D RESULTS jt SS os * y% J ATA"."TPCDS_SF1OOTCL". pemnnne? Swenennnee” "CALL_CENTER"; Metadata Results Warehouse Long-term Cache Cache SSD Cache Centralized Storage “Local” “Remote” (QUERY RESULTS ARE REUSED FALL OF THE FOLLOWING CONDITIONS ARE MET SIMENGW QUERY SYMTACICALY MATCHES IME PREVIOUSLY EXECUTED QUERY. SIME TARE DATA CONTRIBUTING TO THE QUERY RESULT HAS NOT CHANGED. IF THE QUERY WAS A SELECT QUERY, THE ROLE DECUTING THE QUERY MUST KAYE THE NECESSARY ACCESS PRIVILEGES FOR ALL THE TABLES USED INTHE CACHED QUERY, <4 ‘tie aUeRY DOES HOT NCILDE FUNCTIONS THAT MUST BE EVALUATED AT EXEGUTON TE EG. CURFENTTMESTAMP AND UUID_STHING. (NOT THAT THe THE CURRENT_DATE FUNCTIONISAN SR Se ie ata aca Gace aos oe eae nea a) °Y QO © ) SELECT CC_CALL_CENTER_SK AS cc FROM “SINOWFLAKE_SAMPLE_D ATA""TPCDS_SFIOOTCL". "CALL_CENTER"; RESULTS. Metadata Results Warehouse Long-term Cache Cache SSD Cache Centralized Storage “Local” “Remote” C Warehouse Long-term SSD Cache Centralized Storage “Local” 4 ” eae EE eK ES) eetnay Rome tot BUS) UNDERLYING DATA G SECTION 10 — SNOWPIPE © SNOWPIPE INTRODUCTION FIREHOSE 7 Snowpipe Service External Bor GO $3 notification $3 Ts File data saber SNOWPIPE Snowflake Database = = Y S% SECTION 11 - LOAD DATA FROM AZURE © AZURE COMPONENTS ¢ AZURE INFRASTRUCTURE HIERARCHY Ly u LOAD DATA FROM AZURE © CREATE AZURE FREE TRIAL ACCOUNT © SETUP DATA LAKE AND ITS OBJECTS * CONFIGURE SNOWFLAKE TO INJECT DATA FROM AZURE INTO TABLES. + LOAD CSV DATA FROM AZURE * LOAD PARQUET DATA FROM AZURE YJ Sy J ~ AZURE INFRASTRUCTURE HIERARCHY = C7™. eo SECTION 12 — SNOWFLAKE TABLES © TYPES OF TABLES AVAILABLE IN SNOWFLAKE © PERMANENT TABLES © TRANSIENT TABLES ¢ TEMPORARY TABLES I ida eS a TYPES OF OBJECTS IN TD SNOWFLAKE

    You might also like