0:00Hey, it's Tim here and this is the third
0:01video in my series talking about Tableau
0:04file types.
0:04Today we're talking about the Tableau data
0:06extract. Now this is actually an
0:07interesting
0:08file type because it's had a little bit of
0:09a change over the last couple of years,
0:11we're going to cover that in a little bit,
0:13but otherwise you can really improve the
0:15quality of
0:15your workflow working in Tableau if you
0:17fully understand what extracts are, how to
0:19create them,
0:20some of the benefits of having an extract
0:22and also some nuances to watch out when
0:24creating an extract.
0:25Add to that the recent changes to the Table
0:27au data model and there's also a few new
0:29things
0:29that you might not be aware of with Tableau
0:31extracts that I'm also going to cover in
0:33this
0:33video. Okay let's get stuck in. Okay so in
0:35front of me right now I have four different
0:38file types,
0:38so I've already done a video on a TDS and
0:40the TDSX, these are these two files that I
0:43'm clicking
0:43right now. We did a video on Tableau book
0:46marks as well but the Tableau data extracts
0:49or otherwise
0:49called a TDE in the past or hyper in the
0:52present world is over here on the right
0:54hand side,
0:54it's actually this icon here with a little
0:57sort of database icon with an arrow going
0:59from one to the
1:00other. Now it's a really important file
1:02type to be aware of because if you're
1:04working with data sets
1:05such as databases or really cumbersome
1:07excel files, when you create an extract it
1:09can actually
1:10improve the performance of what you're
1:12doing in Tableau. So what exactly does an
1:14extract do,
1:15what are the benefits and what are they?
1:17Well as ever as I pretty much do in every
1:19single video,
1:20I don't just sort of talk to you from
1:21experience, I refer to the documentation.
1:23Tableau documentation
1:24is really really good in this field. So the
1:27Tableau data extracts has several benefits.
1:29Number one if you're working with large
1:31files it can optimize those large files and
1:33make it
1:33much faster and easier to work with that
1:35data. Number two it lets you take a
1:37snapshot of the
1:38data set that you're working with so that
1:40you can essentially work on a smaller set
1:42of the data
1:43or let's say you're connecting to a
1:45database that is maybe far away
1:47geographically or is very slow
1:49or maybe is changing because it's still in
1:50development, you can take a snapshot at a
1:52point
1:53in time save that locally on your laptop
1:55and so that allows you to carry on with
1:57your build process
1:58and develop your workbook but then when you
2:00're ready you can then switch back to a live
2:02connection
2:02to the database and improve the quality of
2:04what you're doing without having to wait
2:06for things
2:06to get into sort of a stable state. The
2:08other benefit is that occasionally there
2:10are some
2:11computations that are much faster done
2:13inside of an extract than natively inside
2:15of a database.
2:16This is a sort of very nuanced aspect we
2:18won't go into too much of that into this
2:20video but I
2:20will highlight to you where this is
2:22highlighted by Tableau and how to find out
2:24those instances
2:25in use cases. The very last benefit is of
2:28course an extract allows you to take a
2:30snapshot which
2:31you can then of course share with someone
2:33so if you maybe want to take a snapshot of
2:35data to give
2:36to let's say a consultant who is coming on
2:38site for just one day you can give them an
2:40extract
2:40they can build off that extract and then
2:41when you're ready you can connect it back
2:43to your
2:43database many days or many months after
2:45they've gone and delivered the work. So
2:48those are some
2:48of the high level benefits and those
2:50benefits are sometimes not really worth
2:52considering when you
2:54create a new workbook or start out with a
2:56new file because fundamentally my viewpoint
3:00is if you want
3:00the best performance typically always work
3:02with an extract unless you're connecting to
3:04a published
3:04data source or you're working with a
3:06database that is designed to deliver those
3:08speeds. One thing you
3:10can then do is get those extracts to update
3:12and refresh so let's get stuck into Tableau
3:14I've talked
3:15way too much let's dive into the product
3:17and find out how it all works. So I'm going
3:19to switch over
3:20to Tableau here I've got a very simple
3:22Tableau instance open here and what I'm
3:24going to do is
3:25I'm going to connect to the superstore file
3:27now this is a very standard file in Tableau
3:29you can
3:29also connect to a database if you're
3:31following along that's perfectly fine too
3:32this will work
3:33exactly the same way. I'm going to go and
3:35select the sample superstore the second
3:37saved data source
3:38here and I'm just going to click on that on
3:40the bottom left here and it opens up a new
3:42Tableau
3:42instance now this is great because what
3:44this does is it connects us to the excel
3:46file that's located
3:47in our documents folder which has all the
3:49data sources that Tableau uses so if I just
3:51go here to
3:52properties you'll see that it's actually an
3:54excel file that I'm connected to and I know
3:56that this is
3:57a live connection because if I just focus
3:59in on this cylinder here at the top because
4:02it's just
4:03the cylinder and nothing else I know that
4:04this is a live connection. Let's go ahead
4:07and create an
4:07extract so to do that we just right click
4:10on the data source and as we go down you'll
4:13see this
4:13option here to extract the data go ahead
4:15and click on that and then you get this
4:18option. Now I'm going
4:19to come back to this page in a second what
4:21I want to do is just show you the process
4:23of creating an
4:24extract then I'll explain each of these
4:25features here on this page in more detail
4:27so let's carry
4:28on through leave everything as it is all
4:30the defaults let's just create an extract
4:32now the
4:32second thing it's going to ask you is where
4:34do you want to save that extract so the
4:36extract this time
4:37I'd like to save it in my Tableau file type
4:40folder and I'll call this version two so
4:42that it
4:42doesn't overwrite the previous one and I'll
4:45go ahead and click save and normally
4:46depending on
4:47the size of the file you might not even see
4:49a window saying that the extract is being
4:51created
4:51so really small files that will just happen
4:53instantaneously the way you know that it's
4:55now
4:56an extract is twofold number one you'll see
4:58this icon here on the top left has changed
5:01also if you
5:01right click on this data source you now get
5:04the options to change whether this is using
5:07an extract
5:07or not so you can see here that before
5:09these options were grayed out and now I can
5:11actually
5:12use them and there's a big tick next to the
5:14use extract option and there's also some
5:17options now
5:17available for me because I'm working with
5:19an extract one of those being a refresh and
5:22also
5:22the ability to understand the history and
5:24the properties of the extracts again I won
5:26't dive
5:26into too much of that into this video I'm
5:28just trying to show you how it works that's
5:30basically
5:31it that's an that's an extract in a
5:32nutshell let's go ahead and find the
5:34extract and see what happens
5:35when you open it so let's just go back here
5:37to my file browser and you'll see here that
5:40I have the
5:41two files that I have previously I have the
5:43original one that we had in the folder then
5:45the
5:45one that I just created called v2 and you
5:47will notice here that it has an extension
5:49called dot
5:50hyper I'll come back to this very very
5:52shortly but now that we have the extract
5:53here if I just
5:54double click the one that I created what
5:58should happen is it will open up a new copy
6:01of tableau
6:02and it will already have the data source
6:04already connected now interestingly here
6:06our extract
6:07although it's been created it's just a data
6:09source so it essentially captured all the
6:11information
6:11about the data and so what I still need to
6:14do here is bring it in and there we go we
6:16pretty
6:16much have a data source ready to go now if
6:19I just close this and let me just open the
6:21previous one
6:21you'll see some slightly different behavior
6:23so let's just open up this first one that I
6:26created
6:26and let's just double click that this is
6:29actually an eu superstore file rather than
6:32the american
6:33one and when I double click this one again
6:35tableau opens up but you'll notice that it
6:37has
6:38a data model here that says extract so the
6:40key thing to remember here is that because
6:44of the
6:44additions to the data model extracts have
6:46become a little bit more complicated but
6:49also it's still
6:49as simple as it used to be I know it's a
6:51slightly sort of weird way of talking about
6:53it but if I
6:53open up this sort of logical layer here and
6:55go into the physical layer you'll see that
6:57it's
6:58still an extract and it's all the same
6:59information it's nothing really sort of
7:01that different now if
7:03I close this as well and I go back and I
7:06just drag one of these hyper files into the
7:09canvas you will
7:10notice a different behavior which is it
7:12will add this connection to the existing
7:14workbook so it
7:15will open up the connection window if I go
7:17back to sheet one you will see that I'm now
7:19connected
7:20to my extract it's got the single cylinder
7:22okay and this is always a really confusing
7:25aspect of
7:25tableau if you connect to an extract well
7:28tableau thinks you're connecting to a data
7:30source of course
7:31it shows it as a live connection to the
7:33extract so if I right click on here and I
7:36go to properties
7:37you'll see this is actually a hyper file
7:39but it treats it like a live connection so
7:41sometimes that
7:41can be confusing you're working with an
7:43extract but you've only got the single
7:45cylinder icon that's
7:45just something sort of to get your head
7:47around and to be aware of you can then of
7:49course take
7:50an extract of the extract it just gets
7:51really complicated very quickly so I think
7:54that could
7:54be improved but just bear in mind if you
7:56want to work with an extract you just drag
7:58it in and it
7:59will come into the existing workbook that
8:00you've got open now let me just close this
8:02connection
8:03and go back to the file types now you'll
8:05see here that it says sample e superstore
8:08dot hyper or
8:10sample superstore v2 dot hyper now
8:12depending on the version of tableau that
8:14you're using
8:15this actually used to be called something
8:16slightly different it used to be called a
8:18tableau data
8:19extract a dot tde file and in tableau 10.5
8:23tableau introduced the new file format for
8:26hyper
8:27and the key thing here is if you're using a
8:29version that is older or is 2020.3 sorry
8:34newer
8:35or is 2020.3 then essentially what will
8:38happen is that every time you open a dot t
8:40de extract or
8:42tableau data extract with a dot tde format
8:45it will be automatically upgraded to a
8:47hyper
8:48file so let me just show you the
8:49documentation on this particular capability
8:51you'll see here
8:52that if I just go back specifically in 2020
8:56.3 extracts will get upgraded to the new
8:59format
8:59they're essentially trying to deprecate the
9:01dot tde format so appreciate a lot of
9:03people watching
9:04these videos are sometimes using older
9:06versions of tableau and so that's just
9:08something to be aware
9:09of now that said pretty much most of the
9:11features of an extract still exist between
9:15a tde and a dot
9:16hyper they're essentially the same thing I
9:18've actually got a podcast with a friend r
9:20avi mystery
9:21and also zen master I'll put a link in the
9:23description or pop up on screen at the
9:25moment
9:25but essentially you get all the same
9:27functionality you don't need to think too
9:29much about the
9:30differences between them unless you're
9:32using something very specific for example
9:35if you use
9:35a hyper file you'll get better performance
9:37you'll see much more improved capabilities
9:40and so
9:40it's actually a desirable format to be
9:43using rarely with tdes do you actually want
9:46to be
9:46using them if you've got the latest and
9:48best of tableau okay so let's just carry on
9:50with the video
9:51and I'll just sort of keep talking as if
9:53hypers and tdes are the same thing if I
9:55mention hyper
9:56just loosely be aware that it will
9:58generally work with the tde file as well
10:01and if we get into
10:02nuances this video will get too long so
10:05that's pretty much the sort of the bare
10:07bones of an
10:08extract now what I didn't do is explain
10:10some of the options around extract so if I
10:13just go back to
10:14our file you'll see that I've got the
10:16extract here on my desktop and if I go back
10:18and I remove the
10:20extract and there's two ways to do this you
10:22can just untick this option here that says
10:24use extract
10:25and when you do that the cylinder switches
10:27to a single cylinder right so you get that
10:30double icon cylinder removed and you go
10:31back to a single this essentially means
10:33that instead of
10:34pointing to my dot hyper file in my folder
10:37here it's actually pointing back to the
10:40excel file
10:41okay and then if I go back to this view and
10:44I right click and I say use the extract
10:46it switches back to the hyper file so you
10:48can essentially switch between a live
10:51connection
10:51and the snapshot as it were or the snapshot
10:54and the database whatever it is you can
10:56switch between
10:57those just to make sure everything is
10:58working really really well this is great if
11:00you're
11:00building something and it's maybe
11:02development or prod and you can just switch
11:04to the live database
11:05to make sure that the data is coming
11:07through properly maybe it doesn't work I
11:09find the switch
11:10it back to an extract and you're all good
11:12okay so that's pretty much it so what I
11:13actually want to
11:14do now is remove the extract I want to get
11:16rid of the extract altogether so actually
11:19if I go here to
11:19the extract option and I go remove what tab
11:22li will do is it will just remove the
11:24extract okay
11:25and you can also ask you to delete the
11:26extract so this will go ahead and delete
11:28the file so let's go
11:29ahead and do that click okay so now it's
11:31removed the extract it's deleted it if I go
11:34back to the
11:34file folder you'll see here that the v2
11:37file has disappeared so it's sort of
11:40cleaning up after
11:41itself when it's removed the extract which
11:42is a great sort of capability now the
11:44reason I did
11:45that is because now I want to create an
11:47extract and go through some of those
11:49details in more
11:49detail so let's go in here and let's go to
11:52extract data and let's look at the first
11:56key thing that
11:56you're faced with which is the logical
11:58tables at the top and physical tables now
12:00if you're not
12:01familiar with these terms highly advise you
12:04check out my video on the data model I've
12:06done really
12:07sort of concise summary of what the data
12:09model is and in that video I talk about
12:11logical and physical
12:13layers there's a terminology that tableau
12:15uses to help explain how the data model
12:17works and
12:19fundamentally the data model brings a few
12:21complexities to extracts because up until
12:24the data model 2020.2 there used to be a
12:27capability to do multi-file extracts okay
12:30and
12:31multi-file extracts it's actually probably
12:33better just to show you what I mean okay so
12:35let's let's
12:36make this simpler so at the moment you can
12:38see the logical tables option is ticked
12:39there is no
12:41physical tables options capable to me okay
12:43so there's nothing I can do to tick this so
12:46again watch my data model video to
12:48understand more on that so let's go in here
12:50and edit the
12:51data source and what you'll see is that I
12:53have orders people and returns so this is
12:56just a
12:56standard data model okay nothing too
12:59complex and the reason this is behaving
13:01this way is because
13:02I haven't got any sort of physical tables
13:05beyond these logical tables that need to be
13:08worked in so
13:09tablet looks at the data source and
13:10realizes hey the only thing you need here
13:12are logical tables
13:13orders peoples and returns so if I go into
13:16orders and I open that up and then what I
13:19do in here is
13:20I create a join on returns it's going to
13:23complain a little bit so let's just say
13:25order id and
13:27this should be returns yeah let's just let
13:32's just create create it like that and just
13:34do a left join
13:35I think the returns data set is a little
13:38bit weird but we'll just close this and we
13:40'll leave it like
13:41that so here we are we've got the basic
13:43information here about the returns and now
13:46I go back close
13:47that and then go back to my sheet that
13:50change has taken place it's still a live
13:52connection
13:53but now if I right click go to extract data
13:56you'll see that I now get the option for a
13:58physical table
13:59okay so that's the key thing to be aware of
14:02okay the physical table option will only
14:04appear
14:05when you actually have physical tables and
14:08joins inside of the logical layer as it
14:11were okay so it's
14:12a slight sort of complexity to be
14:14complexity to be aware of again if you
14:16check out the documentation
14:17it goes into examples and instances where
14:19this is going to specifically happen a good
14:21use case is for
14:23example row level security when you do
14:25probably actually want a physical table so
14:28that the join
14:29is done in real time rather than being pre-
14:31computed and blowing up your data set into
14:34this sort of
14:34really tall sort of duplicated data set
14:37okay so that's probably an example where
14:40you want to be
14:40able to do that if what I just said sounds
14:43like alien whatever just don't worry about
14:46it and
14:46honestly there's lots of blog posts you can
14:48google and when you come across the feature
14:50you worry
14:50about it then but just remember to enable
14:53these physical table option you need to
14:55have a physical
14:56table inside of your data model okay but I
14:58'll stick to logical tables and I'll go to
15:01the next
15:01option which is filters so filters allow
15:04you to do exactly that they allow you to
15:06filter your data set
15:08before you take the extract okay so
15:10extracts are an important part of the flow
15:13if you think about
15:14the order of operations for example and
15:17where if you exclude a bit of information
15:19from your extract
15:20then it's not going to be available from
15:22that point onwards if you're working with
15:24that extract
15:25so if I go in here and I add a city extract
15:27and I just say that I'm only looking at
15:30Aberdeen
15:31and I click at okay I'm working with the
15:33European data set you'll see here that it's
15:35the city keeps Aberdeen so when I create my
15:37extract it will remove this from my
15:40snapshot
15:41so my snapshot will shrink in size and I'll
15:43only be working with data from the city of
15:46Aberdeen
15:46okay you can add multiple conditions as
15:49well so maybe I only want to look at
15:52Aberdeen and I only
15:53want to look at furniture cells in Aberdeen
15:56so you can see that I'm now building sort
15:58of a nice sort
15:59of set of conditions and this is great
16:01especially if you're building like an ad
16:03hoc report you just
16:04need to filter a large database down to
16:06maybe a small customer or particular area
16:09and so you take
16:10an extract you filter it down and then you
16:12just build with that locally on your
16:13machine okay so
16:15that's basically what the filter option
16:17does the next thing is aggregation okay so
16:20let's just take
16:21this extract now and then we'll take a look
16:23at it and see what happens when we change
16:24the aggregation
16:25option so let's just take an extract again
16:27it's going to ask us where to save that we
16:29'll just call
16:30this Aberdeen furniture I really hope we
16:33have furniture being sold in Aberdeen
16:37otherwise this
16:38is going to return next to no data in my
16:40extract so there we go it's done the
16:42extract and surprise
16:43surprise there's no furniture being sold in
16:46Aberdeen what I'll do is I'll fix that by
16:48removing the extract we'll do this again
16:51for the benefit of the demo and we'll go
16:54back in
16:54we'll extract the data and this time we'll
16:57just add the city of Aberdeen as an example
17:00okay so
17:01we know Aberdeen exists there's definitely
17:04cells we'll do extracts again and call
17:06Aberdeen only
17:07hit save and now if I go look at the
17:12summary of the data I have one row of data
17:16okay so this is a
17:18very sort of small subset of data but this
17:21is the only data available in the whole
17:24entire
17:25workbox just single right if I then remove
17:28that again let's remove and delete that
17:30again and let's
17:31just change this extract data I'm just
17:33repeating this for the benefit of the video
17:36and let's say
17:37this time for the category I'm only going
17:39to analyze technology cells okay let's
17:41click okay
17:42click extract and then call this technology
17:46okay now that's gone ahead and done that
17:50and if I go
17:51and do the preview you'll get a lot more
17:53rows 1847 but notice that every single one
17:56here category is
17:58just technology if you don't believe me let
18:01's drag the category item from the product
18:04hierarchy and
18:05the only thing we'll get back is technology
18:07now watch what happens when I switch back
18:09to the live
18:10connection by unticking this use extract
18:13you see the other categories come in
18:15because in my main
18:16data source of course I haven't filtered
18:18this I haven't filtered this from the data
18:19source you
18:20do that with a data source filter but we
18:22haven't done that here so if I switch back
18:24to my extract
18:25it's just looking at a subset of data so
18:27now you can see how extracts are useful
18:29especially when
18:30you're creating snapshots okay we've only
18:32got a couple more things to cover off the
18:34next thing I
18:34want to show you is the aggregation option
18:37so let's just right click on this and if I
18:40make sure I'm
18:40actually in here and then remove the
18:42extract again we're going to delete the
18:44extracts while we do this
18:46right click and let's go to extract and you
18:49'll see that we have the option that we've
18:51had all along
18:52I'm going to go ahead and add the same
18:55category and technology section just so we
18:57can sort of
18:58build on our previous understanding but now
19:00you'll see this option to aggregate data to
19:02visible
19:02dimensions so what does this actually mean
19:04essentially what tablet will do is it will
19:06analyze your data set and figure out which
19:08dimensions are being used and which ones
19:09aren't
19:10then having done that it will essentially
19:12try and group the data up so instead of
19:14maybe saying having
19:15a million rows at the day level or at the
19:17time level you know and the level of detail
19:19that you're
19:20not analyzing it will look at the data see
19:22what you are analyzing and try and reduce
19:24the number
19:24of rows by adding up the rows where
19:26possible so let's say you're counting the
19:28number of cells or
19:29the amount of profit it will aggregate that
19:31up so that you just get maybe one line
19:33where there
19:34were previously two lines essentially so it
19:36makes your data set a little bit more
19:37compact
19:38the other option is the ability to choose
19:40which rows you bring in so there is an
19:43option here at
19:43the top called an incremental refresh I'll
19:45come to this in a second but you can also
19:47choose which
19:48subset of data to bring in so maybe you
19:51just want the top you know orders by a
19:53specific sort of
19:54condition so if I say top you can obviously
19:57change this and notice these are actually
20:01based on my
20:02logical data model okay so these options
20:05here are reflecting my logical layers in my
20:08data set not
20:09specifically columns in my data set okay
20:11and I can just say bring in the top 100
20:14rows in the
20:14orders tables okay or I can sample it
20:16instead so instead I can say okay look at
20:19the orders table
20:20and bring in the random sample of 100 rows
20:23or I can bring in a certain percentage of
20:26data depending
20:27on how that works okay so that's a really
20:29really cool feature the last thing is an
20:31incremental
20:32refresh so if I click on incremental
20:34refresh I actually get some new options if
20:36I just bring
20:37this up so that it's all in the screen here
20:40the options I get are essentially going to
20:43ask me
20:43which table I'd like to refresh okay so in
20:45this case previously in the past you never
20:48really had
20:48to worry about this because you had a
20:50physical layer so everything in a physical
20:52layer just gets
20:53updated wholesale essentially it's checking
20:55everything in that single layer to see what
20:57's fresh
20:58in a logical layer you actually have to be
21:00a little bit more specific because the
21:02logical
21:02layer is driving the rest of the behavior
21:05in the data source and in the workbook so
21:07if I select
21:08orders here I then have to tell it how to
21:10identify new rows so in the typical
21:12database this might be
21:14a creation date or modified date or it
21:16might be something like a transaction date
21:18essentially you
21:18always want to make sure there's something
21:20unique every time a transaction is changed
21:22or updated
21:23and then tableau will bring it in and so
21:25that's how you can choose here you'll see
21:27that it
21:28it only lets you choose specific things and
21:30typically these tend to be the measures
21:32okay
21:33and dates again tend to work a lot better
21:35here so if you choose order date it's
21:37basically going to
21:37look for new orders beyond a certain date
21:39we're making the assumption that the order
21:41information
21:42can't be changed once the order is made
21:44okay if there's a return for example that
21:46was actually
21:47created in a separate database right so we
21:49get a fresh set of data even if an order is
21:51returned
21:52you can't edit something or transaction
21:54once the transaction is taking place so
21:56there we have it
21:57if I click incremental refresh by order day
21:59and then choose extract what will happen is
22:02next time
22:02I go to refresh this data it will actually
22:05only update to see new rows so let's just
22:08go ahead and
22:08create this so let's say technology
22:10incremental and click save so now we have
22:16that if I right
22:18click on this and I go to extract you'll
22:20see here that I have the option now to do
22:22an incremental
22:23refresh which will only look for new rows
22:25using the orders table based on order date
22:28or I can do
22:29a full refresh which is essentially the
22:31ability to just expunge the whole lot and
22:33start again these
22:34are the same options you get on tableau
22:36server when you publish this to tableau
22:38server via
22:39workbook or tableau online and essentially
22:41those options sort of persist all the way
22:43through you'll
22:43also see some of those terminology in table
22:45au prep with the ability to update data
22:47sources
22:47incrementally as well okay so that's pretty
22:50much all the options in the extract
22:52capabilities
22:52obviously there's a couple more things we
22:54haven't covered here for example the
22:56history gives you some
22:57sort of context as to how many rows were
22:59updated so if you're constantly updating
23:01the extract and
23:02you start to put a build out history you
23:04can actually see when the last action was
23:06taken
23:07and what happens if I just close this right
23:09click extract refresh incremental it's
23:12going to do that
23:13in the background it will happen so quickly
23:14it doesn't even pop up a dialogue window
23:16but now if
23:16I go back to the properties I have to say
23:19history sorry you'll see that the first
23:21thing I did was a
23:22full refresh and more recently I've done an
23:24incremental refresh here so it's actually
23:27working
23:27really really nicely so you can actually
23:30trace what is new and what's been added in
23:33so you can
23:33see here this time around I added 446 rows
23:36reason being is the previous refresh was a
23:38full refresh
23:39when we had changed sort of the subset of
23:41what was required so it's brought in an
23:42additional 446
23:44based on the way I wrote the condition okay
23:47so that's pretty much extracts in a
23:49nutshell there's
23:50plenty more we can cover here of course I
23:51don't want to make this video too long it's
23:53already
23:53long enough and the best thing to do is to
23:55check out the tablet documentation there
23:57are also a
23:58whole series of great blog posts out there
24:00in the tablet community if you just go to
24:01google type in
24:02tablet extracts what are they you'll find a
24:04whole bunch of great resources but this
24:06video is
24:06primarily just a basic introduction to show
24:09you sort of where to go and these two pages
24:11by tableau
24:12I'm going to put them in the description
24:13are really really fundamental sort of
24:15understanding
24:15if you're working with tableau every single
24:17day and you haven't had the opportunity to
24:19spend some
24:20time really understanding the nooks and cr
24:22annies of extracts then this is a really
24:23good place to go
24:25and I want to call out some specific things
24:27to be aware of so there are sometimes
24:30quirks that you
24:30come across and actually it's to do with
24:32things like how data is interpreted in the
24:34extract
24:35so things like dates and formats things
24:38like when the extract is taken and where
24:40the database is
24:41when the extract is taken can do things
24:43like change dates and so on and so forth
24:45the other
24:45thing is computations some computations are
24:48better and more optimized in an extract
24:50environment
24:51compared to a database and again you really
24:53need to use tableau for some time to really
24:56start to
24:56get to the nuances of how that works okay
24:59and the other thing is you know if you're
25:01working with
25:02this a lot you'll sometimes actually come
25:04up against edge cases where it's actually
25:06better
25:06to do things that are slightly
25:08unconventional so again this this post here
25:11by tableau goes into the
25:12details of that the other thing is extract
25:15upgrade so if you're using a tde and you
25:18are you know
25:19transitioning to a newer version of tableau
25:22it's worth understanding what happens to a
25:24tde
25:24when you start to use it in a new version
25:26of tableau because it is not possible to
25:29downgrade
25:30a workbook from hyper to a tde format so
25:32once you have that extract then it's been
25:34upgraded you
25:35can't go back so what does that actually
25:36mean well if you want to share data maybe
25:38it's a tde file
25:39you update it turns into a hyper file now
25:41you can't share that data source with
25:44someone who was
25:45using a version of tableau that only relied
25:47on tde so that's just something to be aware
25:49of okay
25:50just make sure you don't break your
25:51workflow because there is no way of sort of
25:54going back
25:54and so i'll put these uh blogs or these
25:56articles and documentation in the
25:58description below
25:59and definitely check them out and this is
26:02probably the third video actually now in
26:04the series about
26:04file types in the next video what we'll do
26:06is we'll talk about the most obvious one
26:08which is
26:09the tableau workbook what is a workbook
26:11what's a package workbook and that will
26:13sort of conclude
26:13the series on different file types okay
26:15thanks for watching if you've enjoyed the
26:17video you know what
26:17to do if not let us know in the comments
26:19below and what you'd like to see instead
26:21all right take it
Tableau | Data prep | Data visualisation