YT YouTube2Text

YouTube2Text → Transcripts → Snowpro Core Certification Crash Course - Part 1 - Introduction, Exam Contents & Architecture

Snowpro Core Certification Crash Course - Part 1 - Introduction, Exam Contents & Architecture — Transcript

by Ganapathy Tech Tips · 6,346 words · 908 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:02hello everyone in this video we are
0:05going to discuss about snow proo core
0:08certification exam introduction this is
0:12an introductory video for the series of
0:14videos in which we are going to discuss
0:17about various aspects towards the snow
0:19Pro core certification exam in fact uh
0:22this is the series of videos in which we
0:24are going to cover the contents which
0:27are required for you to pass the snow
0:30Pro course certification exam this is
0:32the introductory video on the specific
0:34Series so you can expect more videos on
0:37the same lines in the near
0:40future moving on uh first thing I just
0:44wanted to uh engage or I just wanted to
0:47assess myself whether am I eligible to
0:50create an video like this in an platform
0:53like YouTube to answer that question I
0:56am having some decent amount of
0:59experience snowflake over the period of
1:02past around 6 plus years I am having
1:05total of around 15 plus years typically
1:07in the data side in the data engineering
1:10data warehousing area on which I am
1:12having close to around six plus years of
1:15experience in Snowflake currently I
1:17holding both the snow proo core Advanced
1:20architect certifications and also
1:23snowpro core certification so I sat for
1:26the snow proo core exam twice actually
1:29um in in fact snow Pro core is available
1:31in the form of an recertification exam
1:34as well so initially we need to sit for
1:36the regular certification exam and then
1:39once in two years we need to renew our
1:41certification at that time we can sit
1:43for the recertification as well so I sat
1:46for the original exam and then again I
1:48sat for the recertification exam and
1:51last year I sat for the snowpro core
1:53Advanced architect exam which in turn
1:56renewed my snowpro core as well so I am
1:59holding the snow Pro core certification
2:01exam for almost around more than 5 years
2:04now actually so uh I'm pretty confident
2:07that I'm having some amount of knowledge
2:09on snowflake so that I can be an
2:12eligible person to take a course like
2:14this on a medium like YouTube and also I
2:17am holding all the Hands-On essential
2:20batches from Snowflake if you are aware
2:22of the snowflake Hands-On Essentials
2:24currently snowflake is offering uh five
2:27various different Hands-On Essentials
2:29batches this is typically an lab
2:31environment where we need to submit our
2:34lab process to snowflake snowflake can
2:37assess it and then they can grade it if
2:39you can get the higher scores and higher
2:41grade snowflake will provide us with the
2:43batches So currently I'm holding all the
2:46handson essential batches it includes
2:47data Arrow data sharing data
2:49applications data L and data engineering
2:52I am having the dcdf framework uh
2:55typical assessment certification as well
2:57with snowflake so I am currently holding
3:00all of these sorts of uh certifications
3:03credentials with me so I am I am somehow
3:06eligible to preach or teach this
3:09specific snowflake certification in
3:12YouTube so you can always reach out to
3:14me via my LinkedIn uh that LinkedIn link
3:16is provided below so if you want to
3:18connect with me you can always reach out
3:20to me via
3:22LinkedIn now we are going to talk about
3:25some of the basic Logistics behind the
3:27snowpro core exam uh how the exam is
3:30going to be uh how it is going to look
3:32like what are all the total number of
3:34questions all these basic fundamentals
3:36we are going to discuss these details
3:38are available to us in the form of the
3:40study guide if you can go through
3:43whatever you are seeing in the screen is
3:44the snowpro course study guide which
3:46holds predominantly all the details on
3:49what you can expect on the exam what are
3:51all the specific snowflake related
3:54things which are tested throughout the
3:55exam what are all some of the
3:57prerequisites which are needed all the
3:59details are available here not much of
4:02this is uh is easy for us to consume for
4:05that reason I took some of the contents
4:06from here and I placed it in the form of
4:09the presentation so that it will be
4:11helpful for people to consume it in a
4:13very easy way now I moving back to my
4:16presentation some of the basic things
4:19which snowflake expects from the people
4:21who are going to be part of that
4:23certification exam so the candidate is
4:25expected to have knowledge on data
4:27loading and transformation in Snowflake
4:29BL must be aware of virtual vrow
4:32performance and concurrency must be
4:34aware of ddl and DML queries this is the
4:36basic fundamental for any data Arrow
4:38engineer we need to understand about
4:40semi structured types of data structured
4:42data semi structured data and structured
4:43data we need to be aware of cloning and
4:45time travel options within snowflake we
4:47need to be aware of data sharing within
4:49Snowflake and we need to be aware of the
4:51snowflake account structure and
4:52management no need to worry we are going
4:54to cover all of these details in a
4:56greater detail in the subsequent videos
4:59who is the target audience for this exam
5:01they are expecting people who are having
5:03minimum 6 months of knowledge in
5:04Snowflake platform prior attempting this
5:06exam so the familiarity with an SQL is
5:09recommended uh in fact whoever is going
5:12to sit for this exam will be from data
5:14varing background so typically these
5:16things are taken for granted and what
5:19are all the prerequisites knowledge uh
5:20this is already mentioned as the part of
5:23the study guide so they are expecting
5:25the basic knowledge on database basic
5:27concepts and basics of cloud
5:29fundamentals so all these terminologies
5:31are pretty common basic terminology
5:34related to databases and SQL uh it's a
5:36pretty common thing uh everybody must be
5:38aware if you are practicing data
5:40warehousing you must be aware of
5:41databases and SQL tables and data types
5:44we need to be aware of what is tables
5:45what is views what are all various
5:47different data types we need to be aware
5:49of selecting and manipulating data which
5:50we call it as ETL typically extract
5:52transform load or elt in the form of
5:54extract load and transform we need to be
5:56aware of use store procedures and
5:57functions we need to be aware of
5:59security how the authentication
6:00authorizations happens within snowflake
6:02it's a typical common uh basic database
6:05Concepts since the snowflake is an sas
6:08offering running on top of Cloud we need
6:11to be aware of some of the cloud details
6:14as well which involves the cloud
6:17computing and its benefits types of
6:19cloud services typically types of cloud
6:21services comes under infrastructure as a
6:23service platform as a service software
6:26as a service lot of options there I'm
6:28having lot of videos on these areas in
6:30my YouTube channel I will upload that
6:32specific videos as the part of this
6:34description link so that you can go
6:36through those videos in a clear way and
6:38then cloud computing architectures we
6:40need to be aware of what is storage and
6:42compute anyway we are going to discuss
6:44in more detail about these things in the
6:46subsequent videos and also in the same
6:48video as
6:49well now again coming to some of the
6:52other logistics related to exam if you
6:54are attempting the snow Pro core exam
6:56for the first time you need to at sit
6:59for the exam which is going to have 100
7:02questions recertification is slightly
7:05different the number of questions will
7:06be reduced in the recertification but
7:08this course is mainly for the first
7:10timers so we can assume uh you as a
7:13person who is trying to attempt this
7:14exam for the first time so you can
7:16expect 100 questions in the exam typical
7:19question times are multiple select and
7:20multiple choice I hope you can
7:23understand multiple select and multiple
7:24choice multiple select is all about
7:26checkboxes so for the given single
7:28question you can expect multiple correct
7:30answers multiple choice is the radio
7:32button type for a given question only
7:34one answer can be correct so you can
7:36expect questions in both these areas so
7:39what is the time limit time limit is 115
7:41minutes which is closely lesser than 5
7:44minutes for 2 hours the languages are
7:46English and Japanese so typically we can
7:48attempt this exam in the English
7:50language and what is the passing
7:52percentage is 750 plus on the scaled
7:54scoring methodology again scaled scoring
7:57concept itself is a completely different
7:59concept I am having a video there I will
8:01attach the link for that video as well
8:03how the scale scoring works so we need
8:05to gain 750 plus out of th000 in order
8:09to pass this exam typically on a way how
8:13we can understand this is 75 percentage
8:16and above in then it's not perfectly
8:18correct uh 75 percentage and above but
8:21we can assume that we need to make 75
8:25questions out of 100 correct then we can
8:27pass this exam validity for this exam is
8:302 years uh pretty sad but this is the
8:33fact snowflake is instructing people to
8:35sit for this exam once in two years
8:38typically you can take the
8:39recertification no need to sit for the
8:40original exam again you can sit for the
8:43Lesser version of exam with lesser
8:44questions and lesser money as well what
8:47is the cost involved it is 175 plus USD
8:49currently and you need to pay the extra
8:51taxes as well it's a pretty slightly on
8:54the higher side of the exam Higher Side
8:57in terms of the cost simple math if you
9:00do simple math 100 questions in 115
9:02minutes approximately we are going to
9:04have 1 minute and 15 seconds for each
9:08and every question or in turns we can
9:10say 75 seconds for one question it can
9:14be doable because the question
9:16complexity is not that much complex we
9:19can attempt this within 75 seconds and
9:22the domain breakdown for the exam uh
9:24this is very important we need to
9:25understand this in a greater way if you
9:27can see this is the breakdown of the
9:28domain which is provided as per the snow
9:30Pro course study guide so predominantly
9:33most of the contents almost around uh if
9:36if you take questions base almost around
9:3825 questions is covered under snowflake
9:40data Cloud Futures and architecture they
9:43are climbing snowflake is climbing thems
9:45as the snowflake data Cloud uh so there
9:47is no different jargon snowflake data
9:49cloud is always a snowflake actually so
9:52account access and security holds 20%
9:54performance concept SCS 15% so since
9:56it's a 100 questioner so we can simply
9:59do and simple math and we can arrive to
10:01number of questions which we can expect
10:03so the key areas which holds more
10:05weightage you need to give more
10:06weightage towards those specific
10:08areas right now um as we saw the
10:12logistics behind all these things um
10:15with respect to what is the exam what is
10:18the total number of questions which
10:19comes in the exam what is the total cost
10:22what is the exam breakdown and
10:23everything so as the introductory video
10:25I don't want to keep it in a logistic
10:27way I just wanted to St start with the
10:29snowflake fundamentals first so in this
10:32video we are going to discuss about
10:34snowflake architecture what exactly the
10:37snowflake is going to offer Us in terms
10:39of the snowflake architecture what are
10:41all the various components within the
10:43snowflake which makes it an unique uh
10:46hybrid Shar nothing and uh Shar data
10:50type of an architecture all these
10:52details we are going to discuss in a
10:54greater detail over this period of time
10:57so this is an uh new uh I cannot say a
11:02new architecture from Snowflake this is
11:04an uh recent one from one from Snowflake
11:07which covers predominant things if you
11:09can see uh typically uh in the initial
11:12days snowflake architecture is slightly
11:14different but this is the abstracted
11:16version of the architecture this is not
11:17the exact snowflake architecture so you
11:20can see snowflake at the bottom which
11:22provides the intelligent infrastructure
11:23elastic Performance Engine optimized
11:25storage all these things which is
11:27provided to us in the form of the snow
11:29flake on top of it uh you can see snow
11:32grid which is the cross Crow offering
11:34from Snowflake and then on top of it you
11:36can run any of your workloads if you
11:38want to run your data aring workload you
11:40can very well run it if you want to
11:42implement a data Lake on top of
11:43snowflake you can very well do that if
11:45you want to implement a uni store on top
11:47of snowflake you can very well do that
11:49again on the abstracted layer you can
11:51see snowflake can solve these sorts of
11:53Industry problems typically the
11:55collaboration the data engineering cyber
11:57security data science and ml if you want
11:59to build the applications on top of
12:01snowflake stream late lot of options are
12:03there so you can build your own
12:04applications as well on top of the
12:05snowflake so snowflake provides the
12:08opportunity to do end to end with
12:10respect to data that is the reason why
12:12they are calling themselves as the
12:13snowflake data Cloud so in the left side
12:16if you can see unstructured semi
12:17structured structured any sorts of data
12:19you can see on the right side you can
12:20see insides predictions monetization
12:22data products and everything it's an
12:24holistic as I mentioned it's an uh 1,00
12:27ft top overview on what are all the
12:30options which are provided to us from
12:32Snowflake this is not the exact
12:34snowflake architecture we are going to
12:35discuss about this in the greater detail
12:38in the subsequent slides now um slightly
12:42to the older versions what are all the
12:45typical um versions of architecture
12:47which we can see first thing is the Shar
12:51dis architecture the as the diagram
12:53represents you can see there is a
12:55centralized place on which all your data
12:57is stored so whatever data you want to
13:00take you need to access it from the
13:02centralized storage location so each and
13:06every memory and CPU associated with it
13:08say you can take this as an individual
13:10computer or an individual compute that
13:12computer can go ahead and it will take
13:15the data from the shared disk so there
13:17are lot of advantages and disadvantages
13:19for the shared disk typically it's an
13:21shared disk storage as we already
13:24discussed the storage medium is shared
13:26across all the computer so it it helps
13:29in data sharing it helps in the file
13:31over aspects as well say if I'm having
13:34this specific server down obviously
13:37there is no problem with my data so
13:39people can very well come ahead and then
13:41they can take the data say tomorrow or
13:43after some time if this one comes up
13:45obviously they can go ahead and take the
13:47data so there is no problem with the
13:49outage so fail over type of the options
13:52are will safeguarded using the Shar disk
13:54architecture other forms of the
13:56architecture is the Shar nothing
13:58architecture this is predominantly you
14:00can see if you are aware of the MPP
14:02engines massively parall processing
14:04engines like teradata uh ntia or
14:07anything they predominantly make use of
14:09the Shar nothing architecture where you
14:11can see distinctly you can see there is
14:14an um typical storage medium which is
14:18attached to each and every server so I'm
14:21I'm talking taking the CPU and memory as
14:23and combined as a server or a computer
14:25so each and every server or node is
14:27having its own usual data disk this is a
14:31shared nothing so meaning this data is
14:33not shared with this specific server
14:36this data is not shared with this
14:38specific server that is what the Shar
14:40nothing is all about this provides lot
14:42of other features what are all the
14:43features it's like you can have the even
14:46data distribution among the nodes this
14:48will help us in achieving the massively
14:50parallel processing uh it's like we are
14:53having some other videos in the same
14:54areas I will attach the links to those
14:57videos as well where we discussed more
14:59about what is MPP how we can achieve MPP
15:02using these sorts of architectures so it
15:04provides better scalability it provides
15:06better performance these are all the key
15:08important advantages of using Shad
15:11nothing architecture okay now snowlake
15:16is trying to combine both of these
15:19worlds typically the Shar nothing and
15:22also the Shar data architecture we want
15:24to utilize the advantages of both the
15:27architectures so snowflake came up with
15:30this specific architecture pattern uh
15:33this is the one which I took from the
15:34snowflake documentation which is widely
15:36available but um to make it a very
15:40simplistic way or to make it in a more
15:42understandable way this is the one which
15:44provides a more detailed things so
15:46typically if you see um as we discussed
15:49earlier it's an hybrid it's an hybrid of
15:52Shar disk and Shar nothing so Shar dis
15:56as we already discussed similar to Shar
15:59dis it provides a centralized data
16:01repository which provides unified access
16:04from compute which is achieved using
16:06this data storage layer at the bottom if
16:09you can see this data storage layer will
16:12provide us the Shar Disk type of an
16:14architecture so the compute layer which
16:17we call it as virtual vrow layer can
16:20connect to the data storage layer and
16:22then they can take the data from it so
16:25it's a centralized layer which is
16:27running on top of any of the cloud so
16:30snowflake as I already mentioned it is
16:32an cloud-based data platform or a data
16:34warehouse which runs on top of
16:37predominant Cloud providers like AWS
16:40Azure or gcp right and Shar nothing
16:44architecture where you can see the
16:47individual virtual warehouses will
16:49provide us the flexibility to take the
16:53data in an individual way that is the
16:55advantage of the Shar nothing which can
16:57be implemented using the compute layer
17:00of snowflake so combining both the wells
17:03together Shar nothing and Shad dis is
17:06achieved using these typical layer
17:09comparison or layer
17:11isolations in three different ways cloud
17:14services is a different layer that is
17:16not uh a major one with respect to the
17:20typical scalability and other aspects
17:23it's it's completely different it takes
17:25care of authentication and everything we
17:26will be going to discuss in more detail
17:28about
17:29right so Shad disk is an centralized
17:32data storage which is provided to us in
17:34the form of the data storage layer on
17:36top of Cloud and then the compute will
17:38be provided in form of the virtual varos
17:41right so by doing so we are trying to
17:43combine both the worlds of Shar dis and
17:46also Shar nothing now what we are going
17:49to do we are going to discuss about each
17:51and every layer in a more detailed way
17:53so that we can understand the
17:56interdependencies between the layers and
17:58also so the individual layer
18:00responsibility within the snowflake so
18:03starting from bottom up uh what is the
18:06storage layer what is the main
18:08significance of storage layer and
18:09everything this is the same one which I
18:11took from the previous diagram so the
18:14storage layer is as simple as it is it
18:16is the layer on which your data is
18:18stored as simple as it is it's just a
18:21storage layer where your data is stored
18:24typically this storage layer will be
18:26laying on top of Cloud aw Azure or gcp
18:30so this this specific data Lake type of
18:33and storage options is provided to us
18:35from AWS via S3 from Azure via blob and
18:38via gcp VIA GCS buckets so snowflake is
18:42utilizing those medium to store the data
18:45now when we load the data into snowflake
18:49say I'm taking an CSV file as an example
18:52what it happens snowflake will
18:55reorganize the data to an internally op
18:58optimized compressed and columnar format
19:01this is very important we need to
19:03understand this in a very clear way so
19:05when we load the data into snowflake
19:08what snowflake engine does it takes the
19:11data and then it do all this
19:14optimization work internally it will
19:16optimize it will compress and then it
19:19will store the data in the columnar
19:21format as you all a oap predominantly
19:25deals with columns so typically we need
19:27to store the data in in the columnar
19:29format then only we can yield the
19:32benefits of oap MPP massively parallel
19:34processing and everything again I am
19:36having an video which talks about the
19:39typical columnar storage I I will be
19:41adding that link as well in the
19:42description so snowflake take the data
19:45it will do its own operations it is
19:48completely a back blackbox uh till now
19:50snowflake didn't reveal what type of
19:52optimization it is doing in what file
19:55format it is storing nothing is revealed
19:57so far so it is taken care completely by
19:59snowflake no need to worry whenever we
20:01push any data into snowflake snowflake
20:03will do the optimization right then data
20:06is stored in Snowflake databases in
20:09always compressed and encrypted format
20:12this is all about data security so the
20:14data is stored within the snowflake in
20:16the encrypted and compressed format
20:18which we already discussed snowflake
20:21automatically organizes the stored data
20:24into micro partitions this is again
20:27another concept we are going to discuss
20:29more about the micro partitions in the
20:30subsequent videos so far for the
20:33understanding you just understand how it
20:34is storing it is doing all compression
20:36optimizations and then it is storing the
20:39data in the form of the micro partitions
20:41again an optimized immutable compressed
20:44columnar format which is encrypted using
20:46aes256 encryption it's an common stuff
20:50as we already discussed encryption and
20:51everything so it is doing it using AES
20:53256 encryption and storing it in micro
20:56partitions micro partitions meaning it's
20:58a partitioning of Big Data it will do
21:01simple simple partitions it will chunk
21:02the data into multiple partitions and it
21:05will store the data in the form of the
21:06micro partitions data is loaded into
21:09snowflake organized by databases schemas
21:11and accessible primarily as tables that
21:14is very important if you can see once we
21:16do all the data loading and everything
21:19how we can access the data in the form
21:20of a table uh so whatever the data
21:22storage medium we will Define the table
21:24layer on top of it say if you can upload
21:27an CSV file into sow flake and then we
21:29can build the file format on top of it
21:31we will build the table so typically
21:33using the an SQL we can go ahead and
21:36querry the data as simple as is so what
21:39are all the various varieties of data
21:40formats which are supported by snowflake
21:42snowflake supports three all typically
21:45all actually it supports all structured
21:47formats it supports all semi-structured
21:49formats it supports all unstructured
21:51formats in all these fions Json AO par
21:55in the document images and audio as well
21:57uh this is a slly trickier one we we
21:59will going to discuss about this uh it
22:01is not actually storing it within the
22:04snowflake uh in fact all the audio
22:06unstructured related documents will be
22:08stored in typical S3 or blob or GCS and
22:12the index and everything will be stored
22:13within snowflake we will be discussing
22:15in more detail in the subsequent uh
22:17videos related to this now moving on to
22:21the compute layer uh computer layer how
22:23it works we discussed about storage
22:25layer typically plain storage when when
22:28we discuss about compute tip this layer
22:32is represented by snowflake as a muscle
22:34layer muscle is all about power uh
22:37whatever the power you need to take the
22:40data out of the shared storage layer is
22:44provided to you in the form of an
22:47virtual varrow again an additional
22:49terminology in the varing concept people
22:53from the data aring background we tend
22:55to say data Arrow as a term data Arrow
22:58is the place where we store all the data
23:01all the data mods and everything but it
23:03is slightly different on the snowflake
23:05terminology whenever we call it as an
23:07virtual wouse we are talking about the
23:10compute layer uh it's an typical compute
23:13which will be provided to us in the name
23:15of the virtual vrow so it's a dynamic
23:17cluster of MPP compute which consisting
23:20of CPU memory and temporary storage so
23:23each and every virtual vrow will come to
23:26us in the form of and server server
23:29meaning in the form of compute with
23:31enough amount of CPU enough amount of
23:33memory enough amount of storage as well
23:36so don't confuse this storage with the
23:38shared storage layer this storage
23:41depends on the size in which we are
23:43going to take the virtual Vos we will be
23:45discussing it all the compu related task
23:48like data loading unloading query
23:50execution data pipeline ml model
23:52training and scoring will happen using
23:55this virtual vrow layer why because the
23:57compute the muszle power is provided to
24:00us from the virtual wouse layer only now
24:04how we can Define how we can take the
24:06virtual warehouses snowflake is offering
24:09the virtual warehouses in the form of
24:10the T-shirt sizes it starts from X small
24:13small medium large extra large till 6X
24:17large so to understand it how it works
24:20say if I'm taking an virtual vrow As an
24:23X small varrow it comes with a single
24:25node which is an8 core bar RS with 16 GB
24:29of RAM and 200 GB of local dis this is
24:33very important so now if I can scale up
24:35to an Next Level say a small you you can
24:38imagine it into two meaning everything
24:41is doubled so if you can take a small
24:43there you will be having 16 cores and
24:46then we will be having 32 GB of RAM and
24:48then we will be having 400 gits of
24:51storage similarly for medium how it goes
24:53you can double it you can double the
24:55small you will get it to medium so it's
24:56an simple doubling EXT to by which you
24:59can get it so as simple as it is
25:01depending on your load depending on your
25:03power how how frequently how speed you
25:07want to get the data out of the
25:08snowflake depends on that you can create
25:11the virtual warehouses so now you can
25:13understand it's a compute layer it's
25:15completely independent of the storage
25:17layer that is the power of snowflake so
25:19if you want to scale compute
25:20individually you can you can scale the
25:22computer individually already the data
25:24storage layer is an independent storage
25:27why because it's already lying on top of
25:29all cloud storage which provides all
25:31sorts of durability all sorts of
25:35availability right so now coming to
25:37again coming back to the compute layer
25:40we can scale it vertically vertical
25:42scaling horizontal scaling it's a common
25:44terminologies within the cloud ecosystem
25:46so typically if you can see scaling
25:48vertically is medium to large say I'm
25:50currently having an medium wouse I want
25:52to scale it to large wouse for more
25:54compute for processing large amount of
25:56volumes and complex queries it's
25:57possible so we can scale it horizontally
26:00as well in the form of the multicluster
26:01warehouses say example I want three
26:04virtual warehouses of large size it is
26:06also possible to meet the demand of
26:08concurrency user concurrency we need to
26:11scale horizontally so some terminology
26:13we need to understand say for more
26:15compute power for more complex queries
26:17we can scale vertically for more user
26:19concurrency we can scale horizontally
26:21right we can pass we can pass and then
26:24we can resume the virtual voses that is
26:26very important so autos and auto resume
26:29options are already enabled by default
26:31within the vrow per second building that
26:34is very important the predominant
26:36snowflake building goes for the virtual
26:38varrow only so it is based on per second
26:40building based on the varrow size using
26:43credits so X small if I take X small can
26:46consume one credit per hour so if you
26:49can take the per second it will consume
26:51not3 credit per second but there is one
26:54kave there for the first 60 seconds it
26:57is built automatically meaning if I'm
26:59starting a Vos now irrespective of my
27:02Vos ter termination or Vos autoing
27:05meaning say I started my vrow I fired a
27:08query it ends with 3 seconds and then I
27:11I can pass my um vrows for consuming
27:15credits but anyway you need to be
27:17charged for 60 seconds the first 60
27:19seconds you need to pay for it and then
27:21only the individual per second building
27:24starts in general if you want to
27:27understand how the building works within
27:28the virtual varrow layer typically three
27:31factors are there number of virtual
27:32voses how long they run and size of the
27:35virtual vrow so we in the SQL
27:38terminology if you can see how we can
27:40create the arrow we can use some queries
27:42like this you can use the RO of s admin
27:44again a different concept we will
27:46discuss more about it in subsequent
27:48videos so create a arrow with the size
27:50of medium Auto suspend after 300 seconds
27:53autor resume equal to True initially
27:55suspended equal to true meaning it will
27:57create virtual vrow in a medium size and
28:01then at the initial level it will be
28:03suspended to true so there is no charge
28:05whenever you fire a query against that
28:08specific virtual vrow then only the
28:10virtual vrow credits will be consumed it
28:13will move into the on State and then it
28:15will process your query and then it will
28:17shut down automatically based on your
28:20auto suspend future right so as we
28:23discussed it comes with all the T-shirt
28:25sizes actually 6X large is also there so
28:28it starts from one and then multiply by
28:30two multiply by two you keep on
28:32multiplying it by two and finally you
28:34will end up with 256 notes actually for
28:376X large which is an very massive
28:40virtual
28:41Barrow right moving on to the next layer
28:44which is the top layer of snowflake we
28:46discussed about base layer we discussed
28:48about the compute layer which is the
28:50muscle power we are going to discuss
28:52about the brain power of snowflake which
28:54is referred as the cloud services layer
28:57typically a brain layer of snowflake why
28:59a brain layer of snowflake it determines
29:02it do all the calculation for you it it
29:05do all the infrastructure management it
29:07do all the query optimization it do all
29:10the security related things a complete
29:13optimization plan or a quy plan is
29:15computed using the cloud services layer
29:17that is the reason why it is called as a
29:19brain layer of snowflake so typically it
29:21does these features of authentication
29:24infrastructure management metadata
29:26management query paring and optimization
29:28access control and encryption these are
29:30all some of the key features which it
29:32does but again as we discuss these are
29:36all needed by snowflake to take the
29:38proper optimal decisions on query
29:41parsing and optimization so the entire
29:43metadata is stored in this layer the
29:46statistics about the data is stored in
29:47this layer so if you can fire a query
29:50against that specific table based on all
29:51those metadata information it will
29:53generate the least expensive plan as as
29:56similar to other engines as as well
29:58right it's a fully managed layer by
30:00snowflake that is very important we are
30:02not going to do anything in fact we
30:04don't have much of a visibility of this
30:06layer as well it's completely managed by
30:08snowflake so whenever a SQL query is
30:11sent to the query Services layer
30:13Optimizer before being sent to the
30:16compute layer for processing right as as
30:18we discussed earlier it does the
30:20optimization work based on the metadata
30:22which is stored there and then it will
30:24send the query to process to the
30:27subsequent layers of virtual vrow layer
30:29and then to the data storage layer so
30:31some important things to see here here
30:33also you can see some caching you can
30:35see the metadata cache you can see the
30:38results cache right so what is the
30:40functionality behind this so results
30:43cache will be storing a cached copy of
30:46your executed query results so if you
30:48run the exact same query within 24 hours
30:51the results cash will be used no wouse
30:54will be required to be active for this
30:56meaning say I'm having a table with 10
30:59rows uh and I youu and select star from
31:02table name on that specific table now at
31:05the initial time it will go to the
31:07virtual vrow it will take the data and
31:09then it will come back within the same
31:1124 hours provided there's no changes
31:13happen to the table if I can fire the
31:16query again select star from table name
31:19it will not go to the virtual varrow
31:21layer itself it will directly throw the
31:23results from the results cache that is
31:25the major advantage here so there is no
31:28consumes consumption of credits within
31:29the vrow layer so your charge will be
31:32greatly reduced and then there is an
31:34another stuff which we call it as the
31:35metadata cache which is very very
31:37important for the layer to perform to do
31:40all the optimization and everything as
31:42you can see it provides the query
31:43optimization data filtering stored in
31:46the services layer improves the compile
31:48time for the queries against the
31:49commonly used tables this is very
31:51important so it's a common stuff there
31:53is no specific relationship to the query
31:56execution but this is the underlying
31:59metadata using which snowflake Services
32:01layer come up with the least expensive
32:03plan results cash is something which is
32:05very important and there is an another
32:07cache which is associated with the
32:09virtual Barrow layer as well that will
32:11also do the same magic not a magic it
32:14will do the again the abstraction say
32:16same query which is fired against it it
32:18will not go to the storage layer instead
32:20if the results are available within that
32:22same layer it will throw it back from
32:24there but again that is closely tightly
32:27integrated with the virtual vrow so if
32:29you can suspend the virtual warrow
32:31obviously that complete data complete
32:34cache will be thrown away again it will
32:36go through so it is completely tied to
32:39the nature of the virtual vrow uh that
32:41query caching within the vrow layer so
32:43these are all the typical caching anyway
32:45we are going to discuss more about
32:46caching in one specific video right last
32:50final thing uh what are all the ways by
32:52which we can connect to snowflake there
32:54are multiple ways by which we can
32:56connect to snowflake typically we use a
32:58web UA option uh web UA is the direct uh
33:01one which using the U URL we can connect
33:04to snowflake uh there is an advanced
33:07version which they calling it as the
33:08snow site these are all the pictures
33:10from snow site which provides lots of
33:11insights with respect to visualization
33:13with respect to the data sharing
33:15everything is there data monitoring data
33:17products everything is available to us
33:18in the form of the single UA uh which is
33:21named as no site anyway we will discuss
33:23more about it in the subsequent videos
33:25right so webu is the one of the major
33:27interfaces which people will be using to
33:29interact with snowflake to fire SQL
33:31queries and everything say if you want
33:33to programmatically connect with
33:34snowflake there is an option called
33:36snowsql where in which you can
33:38programmatically connect you can write
33:39the command line interface command
33:42commands and then you can connect it
33:44from your command line interface say
33:45like your terminal or puty or anything
33:48you can connect to Snowflake and you can
33:50fire a query using the snowsql and
33:52Native connectivity with odbc jdbc cs
33:55are there uh so predominantly you can
33:57see more most of the things are using
33:58odbc jdbc so you can similarly you can
34:01utilize the odbc jdbc native
34:03connectivity to connect to snowflake
34:05there are some native connectors for
34:06Python and Spark as well uh there is an
34:08interesting feature called snow park we
34:10will discuss later about that but you
34:12can natively connect natively you can
34:14connect to snowflake using Python and
34:16Spark connectors and then third party
34:19connectors are very well available
34:20within the snowflake partner ecosystem
34:22like Informatica thir spot lot of
34:24partners are available with snowflake so
34:26using the third party the ecosystem we
34:29can very well connect to the snowflake
34:32all right so these are all some of the
34:34uh Ways by which we can connect to
34:35snowflake so with this we come to end of
34:38this video uh just an introduction about
34:40the snowpro core certification exam and
34:42then we discussed about uh types of uh
34:45architectures and how snowflake
34:47implemented the hybrid Shad nothing and
34:49Shad dis architecture we discussed about
34:51the various layers within the snowflake
34:54architecture like cloud services layer
34:56virtual Barrow layer or compute layer
34:58and then the data storage layer uh
34:59within the cloud medium and then we
35:01discussed about the ways to connect to
35:03snowflake this is the introductory video
35:05for the series of videos we will be
35:07sharing we will be I will be uploading
35:09lot of videos in the same aspects with
35:12respect to the other areas which is
35:14required for the snow proo course
35:15certification in the near future please
35:18do comment uh that is very important for
35:20me to uh enan or add more contents to it
35:24uh please do comment do like or do
35:26dislike as well uh thank you very much
35:28for watching this
35:30video

About this transcript

This page contains the full transcript of Snowpro Core Certification Crash Course - Part 1 - Introduction, Exam Contents & Architecture by Ganapathy Tech Tips, generated from the public captions YouTube serves with the video. The transcript has 6,346 words across 908 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.