數據映
SQL演示

我一直想做的第一件事是探索我可以使用的數據。因此，您會注意到，在數據資源管理器中，我有一個點和單擊的方法來了解我數據湖一部分的不同目錄，數據庫和表格。在這裏，如果我想快速查看不同的表格，我可以概述模式，我可以查看幾行示例數據。現在，當然，所有這些都以開放格式存儲，我可以看到詳細信息，我甚至可以通過單擊按鈕來管理權限。現在，您可以輕鬆地從OKTA，Active Directory或其他身份驗證提供商同步Databricks SQL中的所有用戶和組。很好的是，在此數據資源管理器中，我還可以看到我的三角洲表的曆史記錄，以便我可以理解發生的每個操作。現在，我們將使用開箱即用的樣本NY​​C出租車數據集。但是，還值得注意的是，如果您想從另一個來源設置攝入量，則隻需單擊幾下，就可以與像Fivetran這樣的合作夥伴來做到這一點。好吧，讓我們開始查詢。因此，在這裏，您會看到我們擁有的是您的SQL Workbench。 And this is where you author and execute queries. So let’s go ahead and run this sample one. Now, of course, you can have multiple SQL statements, whether those are ad hoc queries, or DDLs in each of these tabs. I’m also going to go ahead and pin this table since we’re going to use it throughout, while I can kind of go ahead through the schema browser and take a look at some of the other data that I have available to me.

現在，要注意的一件事是，您可以在這裏看到我正在使用每個工作空間的開箱即用的默認SQL端點。但是，這些終點很棒的是，它們消除了簇的複雜性，您不必考慮版本，而不必考慮Spark配置。您隻需選擇一件T恤尺寸，小，中，大。您可以將它們分配給特定的團隊或部門。因此，您可以簡化政府的治理和退款，並且很快就可以使用新的無服務器端點，您無需考慮在短短幾秒鍾內就可以為黃金時段準備的啟動時間。

因此，既然我們已經取得了結果，為什麼我們不看一些執行細節？您在這裏看到我們可以很好地了解實際發生的事情。我們還可以理解我們運行的一些以前的執行，如果需要，請返回以前的版本。If we’re really firing on all cylinders, we can also go ahead and leverage any of these keyboard shortcuts, for things like opening and closing new tabs, or even executing our queries, highlighting certain portions, everything designed to make your analysis as quick and easy as possible. Now, of course, sometimes you don’t just want to generate results, you also want to visualize them and share them with other individuals and teams in the organization. So why don’t we go ahead and add a visualization. And you’ll see that’s right side by side with my with my results, let’s just do a quick scatterplot of the average fare by route, see if we can figure out which which routes we probably want to avoid, because they’re a little pricey. So you can see here, I was able to quickly just add that, that visualization. And I can quickly spot any of these outliers that I might want to avoid when I’m planning my trips – probably better to take the subway. So now this is great, I’m starting to generate some insights already. But what if I want to organize these insights per system and kind of come back to them time and time again with my team? Dashboards are a great way to do this. And of course, I’m sure you’ve guessed by now, we provide a number of sample dashboards out of the box. So let’s go ahead and pick our NYC Taxi one.

Now, you’ll see that this dashboard provides a great example of a number of capabilities you might want to take advantage of, whether that’s conditionally coloring certain values in a table, or adding interactivity via parameters and allowing the consumers of these dashboards to set different date ranges. And what’s great is if there’s a key metric I’m tracking like total trips, I can also right within Databricks SQL set up an alert so I know when certain thresholds are met. So here you’ll notice that I’ve selected my query total trips, I’ve decided I want to know as soon as we crossed that magical 300 mark for that time range. And we just want to get notified once when it happens. So I’m going to go ahead and create the alert. I’m going to tell it to send to myself. And I’m also going to go ahead and decide to make sure the entire team knows too, and make sure that message goes to Slack as well.

好吧，因此，為什麼我們不跳到Tableau，在那裏我已經將其配置為直接從Databricks SQL查詢並點擊這些SQL端點。您可以在這裏看到我能夠在Tableau儀表板上開發同樣的數據觀點。而且，如果我想打開任何東西並進行一些其他分析，例如，我想通過刪除代碼來理解路線票價的差異。您可以很快看到這裏，我可以看到哪些郵政編碼彼此非常接近。您可以看到很短的旅行，非常便宜的旅行。但是，如果我走得更遠，我也可以開始獲得更多的見解，並了解距離實際上更長的地方。這一切都是通過單擊按鈕來針對我的Databricks SQL數據生成實時查詢。因此，希望這使您了解了數據湖上的分析方法是您要使用我們的內置功能還是您喜歡的BI工具的可能性。