Video |

Snowflake | Data engineering | Analytics

Using the History Page to Learn Queries - Snowflake

I'm still learning Snowflake, so let me show you my trick: do it in the interface, then steal the SQL it writes for you.

Tim Ngwena •9 August 2022 •1 min read •7 min watch •2.1K

Part ofGetting Started with Snowflake

Watch on YouTube

Everything you do in the Snowflake interface generates an underlying SQL statement you can inspect and reuse
The Activity > Query History page shows the exact SQL Snowflake ran, including row limits like LIMIT 100
You can copy generated SQL into a worksheet to learn and re-run it, building up reusable scripts
Always highlight the specific line before hitting run to avoid executing your entire worksheet
You can switch between the modern Snowsight interface and the classic console, and save generated SQL to GitHub for reuse

0:00One of the best ways to learn is to be able

0:01to work alongside someone who knows what

0:03they're doing

0:04and can show you how to do something. With

0:06Snowflake, you can control Snowflake either

0:08using the interface or SQL commands, but

0:10with the SQL commands you really have to

0:12have some

0:12experience writing them before you can

0:14understand how they work and how to write

0:16them off the top

0:17of your head. Not many people actually do

0:18that, they use some assistive tools, but

0:20nonetheless

0:21there is actually a nice way that you can

0:23do this just by using the Snowflake

0:24interface,

0:25then looking at the query afterwards and

0:27then viewing the SQL that was written by

0:29Snowflake.

0:30Let me show you how, let's get stuck in.

0:32Okay I'm here in the Snowflake interface, I

0:34'm actually

0:34looking at the new version of the interface

0:36. If this looks very modern compared to what

0:38you're

0:38used to, you can actually switch back to

0:40the old version of the console just by

0:42hitting classic

0:42console. You have to log in again and once

0:44you've logged in again you go back to the

0:46old way of

0:47doing things and if you want to go back to

0:48the modern way of doing things then you can

0:50just go

0:51ahead and hit the snow side option just

0:52here on the top right hand side and it'll

0:54open another

0:55tab. So you can actually have them side by

0:57side, at least for now that's going to be

0:59the case,

0:59in the future that might disappear. But let

1:01me show you what I'm trying to explain. I'm

1:03still

1:04learning Snowflake and I'm still learning

1:06how to write the SQL commands to control

1:08the Snowflake

1:08system and in essence the way I'm sort of

1:10learning to do this is by using the

1:12interface to do things

1:13I would normally do, then seeing what query

1:16was generated to create that activity. Now

1:18this might

1:18not be familiar to lots of people but in

1:20essence everything you do here in the

1:22interface is actually

1:24going to create a query statement. You can

1:25actually look at it. So let's just do

1:27something

1:27very basic. Let's go to my data set here

1:29and you'll see that I have two databases. I

1:32'm just

1:32using the Snowflake trial so if you're

1:34using the trial as well you'll have exactly

1:36the same set of

1:37data. You'll see there's two databases,

1:38there's one called Snowflake and one called

1:40Snowflake

1:41sample data. I'm going to actually open

1:43Snowflake because that's the sort of

1:44central database that

1:46Snowflake used to show you information

1:47about your Snowflake instance so we can

1:49actually kind of look

1:50at the metadata that's sitting behind Snow

1:52flake in this particular area. I'll go in

1:54here to the

1:55information schema and we'll just take a

1:57look at some of the data that we've got

1:58here. We've got

1:59the databases, we've got things like

2:00columns. I'll scroll down to the tables of

2:02this. This is

2:02probably going to be the easiest one to

2:04look at and you'll see that you get this

2:05representation here.

2:06This is in fact just a preview that's been

2:09generated by the interface but if I click

2:11data

2:12preview what that actually does is it kicks

2:14off a query that's running off the compute

2:16warehouse.

2:16You can see it says here compute warehouse

2:18and you get a table. Now what that table

2:20did is it

2:21generated a SQL command so let's go and

2:23look at what that SQL command looks like.

2:25Let's go over

2:25here to activity on the left hand side and

2:27we look at the query history you'll see

2:29there's actually

2:30quite a few things that I've been running

2:31now. This is a brand new trial instance you

2:33can see that

2:34everything I've run has literally just been

2:36done in the last few minutes and so if I go

2:38to the very

2:38top query you can see that I actually get

2:41the query that was run and I get the

2:42specific SQL.

2:43It's been limited to 100 rows but in

2:45essence this is how it's querying from that

2:47particular table.

2:49So I can actually go and take that copy go

2:51to a worksheet. I have a worksheet here

2:54that's

2:54open from a few minutes ago. Let's go ahead

2:56and open that and you can see that I can

2:58actually

2:58just go ahead and paste it there. It

2:59actually had something already in there but

3:01this time I'm going

3:02to load something slightly different and I

3:04'm going to hit run just by highlighting

3:06this. I always get

3:07in the habit of highlighting then hitting

3:09run because what you don't want to do is

3:10make the

3:11mistake of basically running the entire SQL

3:12statement when you're only meant to run a

3:14small

3:15section of it. So I'm just learning to do

3:17this with most tools I'm still sort of

3:19fresh of this.

3:20If there's a better way of doing it let me

3:21know in the comments but as you can see I

3:23ran that query

3:24again ran in 1.2 seconds and you get the

3:26same table. So that's a very sort of simple

3:29way of

3:29doing this. Now okay let's go ahead and

3:31look at this in a slightly different

3:32context. Let's go

3:33back to the home page and what I'm going to

3:35do is I'm going to go to data and I'm going

3:36to go to the

3:37databases option. Now what I'm going to do

3:39is create a new database and we're going to

3:40have a

3:41look at the SQL that this creates. So let's

3:42go ahead and write the name here. So we'll

3:44call this

3:45tableau_tim_test. I can't do hyphens there

3:49so I have to do underscore test. Let's do

3:52that and

3:53we'll call this test_db and once we've done

3:56that we'll hit create. That will go ahead

3:59and create

3:59a database. Of course once the database has

4:01been created there's actually nothing in

4:03that database

4:04so you'll see we just go in there's nothing

4:06there but again we can go to the activity

4:08area look at

4:09the queries and you can see here's the SQL

4:11that was used to create that. Create

4:13database identify

4:14tableau_tim_test comment equals test_db.

4:17That's essentially what was done and so we

4:19can go ahead

4:20and delete the database and we're going to

4:22use the SQL statement to do the exact same

4:24thing again

4:24this time around. So let's go ahead back to

4:27the data area and we can go ahead go to

4:29tableau_tim_test

4:30and what we want to do is drop this. Now

4:32the new SnowSight interface is a little bit

4:34cleaner.

4:34It's sort of I find it a little bit harder

4:36to sort of get around. I prefer the old

4:38interface which

4:39was sort of familiar in some senses. So

4:41what you do is you go to the very top right

4:43hand side here

4:44you see these three dots which are never

4:46descriptive. Go ahead and select drop and

4:48this

4:48will basically go ahead and delete this

4:49database. So remove the database tableau_

4:51tim_test from the

4:52system. A version will be retained in

4:54timetable for a specified amount of time.

4:56Let's go ahead

4:57and drop that and you'll see that it

4:58disappears from this list. That in itself

5:01was a query so we

5:02can go ahead and look at that as well. So

5:03if I just go ahead and refresh this and

5:06just give it

5:06a bit of a second. Yes you'll see that the

5:09drop option came up there. So this was the

5:11SQL to drop

5:11it and this was the SQL to create it. So

5:13let's go ahead I'm going to grab this I'll

5:15rename it

5:16just to make it slightly different. Let's

5:18go to my worksheet go back into the same

5:20worksheet I was

5:20creating before and I'm just going to go

5:22ahead and paste that in there. I'm going to

5:25call this

5:25slightly different I'm going to call this

5:28tableau_tim_test_v2 it doesn't like spaces

5:31and

5:32I'll call this testdb a second time around.

5:35Okay and now that we've done that we're

5:37going to hit

5:38run just on the specific line remember

5:40highlight the line then hit run then it

5:42will go ahead and

5:43it will create the database tableau_tim_

5:45test_v2 successfully created. Let's go to

5:48our databases

5:49and check that that's worked go back go to

5:51data and there you have the test database.

5:54So if you're

5:55learning snowflake this is a really nice

5:57way to have sort of a soft landing into the

5:59technology

6:00but also understand the best way to write

6:02certain SQL statements to get them to work.

6:04Now there will be some instances where you

6:06just have to go and look at the

6:07documentation but I

6:08think this is a good place to start. You

6:10can almost go through during the trial

6:12creating lots of

6:13different sort of walkthroughs of something

6:15and then once you've got that down you can

6:17then look

6:18at your worksheet and look at okay this is

6:19everything that was done to create this

6:21setup

6:22and you can then use that save that in

6:23github or do whatever you need to to sort

6:25of perpetuate it

6:26for the future so you can learn from it. If

6:28you've enjoyed this video let me know in

6:30the comments

6:30below snowflake is slightly different to

6:31what I normally record but hey I'm learning

6:33snowflake so

6:34I thought I'd bring you along on the

6:35journey thanks for watching and I'll catch

6:36you in the next one.

6:38you

6:39[ Silence ]

Rollup

You can learn Snowflake SQL without memorising syntax by performing actions through the interface and then inspecting the exact query Snowflake generated to do it. This turns the interface into a live teacher for writing your own reusable SQL.

Tim is still learning Snowflake and wants a low-friction way to pick up SQL syntax. He uses a Snowflake trial instance with the default sample databases to demonstrate the trick.

The Breakdown

Switch between Snowsight and classic console 0:33

You can toggle between the modern Snowsight interface and the older classic console (via a login and a button), running them side by side if needed. This is just a tip on navigating the UI, not a required step for the main trick.
Perform an action through the interface first 1:27

Do something you'd normally want to script — here, previewing a table's data — using the point-and-click interface rather than writing SQL by hand. Every interface action like this triggers a query behind the scenes, which is the basis for the whole technique.
Find the generated SQL in Query History 2:21

Go to Activity > Query History and you'll see the exact SQL statement Snowflake ran for your action, including details like row limits (e.g. a LIMIT 100 added automatically to previews). This is where you 'steal' the syntax.
Copy it into a worksheet and re-run just that line 2:41

Paste the generated SQL into a worksheet to learn from and reuse it. Always highlight only the specific statement you want before hitting run, otherwise you risk executing everything else in the worksheet by mistake.
Repeat the pattern for other actions 3:30

The same approach works for any interface action — Tim demonstrates it by creating a database via the UI, then dropping it, then checking Query History both times to capture the CREATE DATABASE and DROP DATABASE syntax generated.
Reuse the captured SQL to redo the action manually 5:20

Paste the captured CREATE statement into a worksheet, tweak details (like the database name), and run it to confirm it reproduces the same result you got through the interface — proving you've learned working syntax you can reuse.
Build up a personal library of learned scripts 5:54

As you repeat this across different tasks, save the working SQL snippets somewhere like GitHub so you build a growing, reusable reference of Snowflake syntax you've verified yourself.

Worth Knowing

Interface-generated queries can include limits you didn't ask for, like a LIMIT 100 added to data previews, so don't assume the copied SQL is unrestricted.
Highlighting only the line you want before hitting run matters because running the whole worksheet by accident can execute statements you didn't intend.
Some tasks will still require checking the documentation directly — this trick is a good starting point, not a full substitute for learning SQL properly.

Use It When

Reach for this when you're new to Snowflake (or any tool with a GUI-to-SQL relationship) and want to learn real, working syntax by doing familiar tasks in the interface rather than starting from a blank SQL editor.

How this Rollup was made provenance & method

A Rollup is drafted by AI from the video's transcript, then reviewed and edited by Tim. Everything used to produce this one is listed below — the model, the exact prompt, and the source video — so the process is transparent and reproducible.

Transcription: On-device — NVIDIA Parakeet v3 for recent videos, OpenAI Whisper large-v3 for earlier ones. The transcript never leaves the machine or gets published.
Drafting: Claude Sonnet 5 in the cloud, from that transcript.
Prompt: The exact Rollup prompt (v2) — the full system prompt, unedited.
Source video: Watch on YouTube
Drafted: 5 July 2026 at 09:38
Reviewed & edited: 5 July 2026 at 09:41 · by Tim Ngwena

Model + prompt + video is everything you'd need to recreate a Rollup like this yourself. The one thing we don't share is the transcript.

Rights. The video and its transcript are the property of TN Media Ltd. Unauthorised use or download is prohibited. © TN Media Ltd.