跟讀練習: What Is a Data Mesh? Databricks Unity Catalog Explained - 通過影片學習英語口語

選擇練習模式

每種模式都會開啟全屏練習室。

147 句

如果句子過短或過長，請點選 Edit 進行調整。

to make everything is clean

and it's clear who is offering what what we're gonna do

we are back yeah to the hub a little bit

but this time this hub is marketplace what do we have

in the marketplace is the following everyone gonna announce their products into central place say hey

i offer the sales data hey i offers the hr data hey

i offer the finance data maybe not not arrows sorry for

that yeah arrows okay

so the dotted arrow means not the data gonna move to central place

this is very important to understand the data gonna stay at their places yeah

so it stays at domain a stays at domain b it stays at domain d

and so on it's just we're gonna make one gigantic catalog about the data into central place

so it is a marketplace where everyone announcing their data

i have the sales i have five tables i have 10 tables

so it's like bizarre it is very very cool

because the data analyst they just go to the marketplace

and read it

and they can understand okay the whole company how many data

do we have how many tables how many domains how the size without doing

this journey by visiting each of those domains without running around in circles to find the data

they're to go to one place and my friends she gonna find the data fast

and she gonna understand aha in order to make my reports

i need the sales data i need this i need

that i need from this and i'm gonna be able to work

but this is not only for the analyst as well for

the ai engineer for any new projects have been in the company

we know where the data is um

this is why we're gonna need a central market central place by the way

this is not anymore federated i just introduced you to the data mesh

it's not anymore federated things are going to the middle

so data mesh is no more central team doing the whole thing yeah this is very clear

because we have federated we split the data by domains each domain

they're gonna do the whole thing alone we're gonna have data as a product

that means we're gonna have documentations i'm gonna explain more what is data as a product

because it's way more than that

but we're gonna have each domain

they're gonna make a catalog about their data into central place the marketplace

they're gonna document it they have a product owners

and everything is like systemized and documented with high quality

and all those stuff so the data mesh concept says again physically each domain gonna have

that data they're gonna own the data at their own place

they never go to the central place they never go to the hub

they never go to the central lake house they're gonna stay at

that department at the domain physically gonna stay at the domains

but we're gonna make a virtual place where the whole metadata of the whole company gonna be described

so that if auditor comes they can read the marketplace

if a new developers comes to the company they're gonna go to the marketplace

if any news gate new data use case happen to the company we start from the marketplace

and understand it this picture it is not only the companies

that understood it who understood this as well it's always the data Databricks.

They are very, very up to date with all pains of companies.

This is amazing, really.

I'm not trying here to sell anything, okay?

I'm not sponsored or something from Databricks.

I'm just trying to say here.

Databricks came and understood, oh, companies are moving to this.

How about I'm gonna offer something similar to this?

I would like, like, because Databricks stopped us here.

Let me just take the icon.

Now, I'm breaking a little bit the vendor, okay?

But it is not that important.

Now, Databricks at the start stopped at here and said, hey, you can make

Spark, you can make lake houses.

It's the best thing.

And they start selling this like crazy.

But they saw this is happening.

Companies are making multiple lake houses left and right, and they're going to need something called marketplace.

And now if you are learning Databricks, you're going to understand the origins of the Unity Catalog.

Unity Catalog is made for this is made in order to

make data mesh concepts in order to make domain -driven architectures

this is what we call domain you're gonna hear a lot

if you go as an architect people are making domain oriented

or domain driven architecture

and this is exactly what we mean at the data world we say data mesh

and now the vendor databricks join this picture

this is exactly what they are trying to sell for the companies yeah i have some doubts about it

but it's working

but not for big companies yeah let me just explain where that's why two years ago yeah

i was making my projects

and then suddenly i've seen something called catalog yeah the catalog yeah the unity catalog

they introduced the unity catalog who knows the unity catalog data bricks had something similar like a marketplace

but it was not a marketplace

it was just a database catalog for your project just only for you for your project okay

and they called it the hive metastore it's legacy name you will not find it

but i'm just telling you the story about it they have like some kind

it looks like a database okay you are you are you are looking at the unit catalog

but it was only for you for your projects and They switch the whole thing to a marketplace.

So you see here, it's called my organization.

Well, maybe in big companies, it's going to be like the company's name.

And you will be able to browse not only your projects, but you're going to find as well all the catalogs of the whole company.

Now, I cannot show you the example because I have here only one example.

So imagine here, I'm going to see inside my organization, maybe 30 catalogs, the sales, after sales, the manufacturing and all those stuff so

that means they made this local catalog

that is available only for a project they made it only visible for the whole company

and this is what suddenly i was like working towards

and this happens to my project i was working at my project

and i saw only one catalog this is my catalog

and i'm doing the server bronze gold and i'm building the warehouse

and then in the next day i was not following the news i opened my catalog

100

and i saw 200 catalogs that's all because suddenly i'm able to see the whole company catalog.

101

They turned each local instance or each local project into like a global catalog that is available for everyone.

102

So everyone was able to see my projects and I am able to see everyone else's projects.

103

So that means if you are working in Databricks, this is the Unity catalog, my friends.

104

And with that, we just, I don't know, I think they did cost them a lot of work they just turned the local catalogs

105

that each project have into a global one

106

and they solved this issue with the data mesh

107

and that's exactly why we have now something called unity catalog

108

this is exactly the reason behind it they want everyone to be able to access everyone

109

or let's say everything should be visible for everyone else this is still issue the thing is the vendor, they are trying to, yeah,

110

to sell us this idea, which is amazing, I think.

111

But this is going to work, yeah, I'm going to say only if everyone is working in Databricks, right?

112

So if everyone is in Databricks, this is going to work because they're going to share their stuff to the middle, right?

113

Now, if a company gets, I don't know, sophisticated and they do things outside of catalog,

114

this is going to be challenging to make the Unity Catalog as a central place.

115

They try it's like there's some work around and stuff

116

but they it could work of course but

117

if the company have like on -prem solutions

118

if one domain is not working at all in the cloud

119

if other is using i don't know yeah azure we have azure stuff fabric

120

if the company is not clean about that platform strategy

121

and they allow multiple platforms then sadly this will not work that's why in some companies,

122

if they are big enough to say Databricks is not enough for them, they have to make their own catalog.

123

Like at the end, I'm going to show you a few examples.

124

I have made some with Claude.

125

I'm just prepared a few visuals for you.

126

Yeah, I don't have your catalog icon.

127

Let's bring this here.

128

So the company should then develop their own marketplace for this.

129

So my friends, this is very cool to understand what is a data mesh

130

and how our data is turning into something called products

131

so it is not anymore this picture where you have central data engineers just prepare the data

132

and give it the world is getting bigger

133

we are getting a lot of new things

134

we are focusing on only on domains we have yeah the product owners

135

we have a marketplace for data yeah

136

we are transitioning to this domain driven architecture

137

i'm really happy with this yeah it's like one of the best i'm gonna say

138

so far best practices on how to do it i love

139

this idea to have the metadata data into stored into one place

140

and as well to have the domains spread

141

it around don't store anything centrally for big companies it will not work

142

and yeah those discussions again happens at the elite level of the architect the principal architects

143

and let's say the big managers the data officers and stuff like

144

that so it's not something that we decide at our level level

145

but if you hear about it if you join a company

146

or something that they are doing domain you should know about the structure

147

and it is relevant for you as a data analyst and stuff like that.

下載應用

流利口語所需的一切

AI 發音評分逐句打分

音標練習掌握每個音

詞彙學習建立你的詞庫

詞彙遊戲邊玩邊學

掃碼

Download on theApp Store Get it onGoogle Play

同頻道

How Modern Data Teams Work (Engineer, Analyst, Scientist, Architect, ML)

Daily ConversationUniversity Students

Learn Databricks for FREE (Step-by-Step Guide)

Business EnglishUniversity StudentsWorking Adults

FOR YOU

本課的詞彙與口語要點

這節口語課以影片“What Is a Data Mesh? Databricks Unity Catalog Explained”為素材。影片中反覆出現的詞有：catalog, domain, marketplace, project, databricks。這段影片共有 147 個句子、1755 個單詞可供跟讀。講話部分時長為 9:10。說話人語速較快，每分鐘約 191 個詞，會出現較多連讀和弱讀。 85% 的單詞屬於英語最常用的 3,000 詞，其餘的詞建議在練習前先學一下。

影片中的重點詞彙

影片中 15 個值得學習的單詞，附發音和釋義：

單詞	發音	釋義
catalogue 名詞	/ˈkæt.əˌlɔɡ/	目錄 /目錄
domain 名詞	/ˌdəʊˈmeɪ̯n/	領域 /領域
marketplace 名詞		草市, 市井
unity 名詞	/ˈjunɪti/	統一 /統一
mesh 名詞	/mɛʃ/	網孔 /網孔
architecture 名詞	/ˈɑː.kɪˌtɛk.t͡ʃə/	建築 /建築, 建築學 /建築學
arrow 名詞	/ˈæɹəʊ/	җян, 箭
architect 名詞	/ˈɑɹ.kɪˌtɛkt/	締造者 /締造者, 創造者 /創造者
vendor 名詞	/ˈvɛn.dɚ/	賣主 /賣主, 賣家 /賣家
visible 形容詞	/ˈvɪzɪb(ə)l/	看得見 /看得見, 可看見的 /可看見的
introduce 動詞	/ˌɪntɹəˈdus/	介紹 /介紹, 紹介 /紹介
database 名詞	/ˈdeɪtəˌbeɪs/	資料庫 /資料庫
announce 動詞	/əˈnaʊn(t)s/	通告, 公佈
icon 名詞	/ˈaɪ.kɑn/	聖像 /聖像
gate 名詞	/ˈɡeɪ̯t/	門口 /門口, 門 /門

值得反覆跟讀的句子

影片中簡短而完整的句子，可以直接用在日常對話裡：

I'm not trying here to sell anything, okay?
I'm just trying to say here.
How about I'm gonna offer something similar to this?
I'm just prepared a few visuals for you.

影片中的語法

說話人最常用的結構，並附上影片中的原話：

結構	影片中的用法
現在完成時 have/has + 過去分詞 — 過去發生但與現在仍有關聯的事	have been · have federated · i've seen
被動語態 be + 過去分詞 — 強調發生了什麼，而不是誰做的	be described · is made

需要注意的發音

說話人用了 68 次縮略和弱讀形式，例如 gonna, I'm, they're。請按聽到的簡短形式來說。

“sh” 和 “zh” 音: mesh /mɛʃ/, transition /tɹænˈzɪʃ.ən/, documentation /ˌdɑkjəmɛnˈteɪʃən/, azure /ˈæʒɚ/
長單詞——注意重音位置: architecture /ˈɑː.kɪˌtɛk.t͡ʃə/, developer /dɪˈvɛləpɚ/, documentation /ˌdɑkjəmɛnˈteɪʃən/, metadata /ˈmɛt.əˌdeɪ.tə/

中文母語者容易讀錯的音：

/v/ — 不要讀成 /w/，上齒輕觸下唇: vendor /ˈvɛn.dɚ/, visible /ˈvɪzɪb(ə)l/, server /ˈsɝvɚ/, solve /sɒlv/, virtual /ˈvɜɹt͡ʃuəl/
詞尾子音 — 要讀清楚，後面不要加母音: catalogue /ˈkæt.əˌlɔɡ/, mesh /mɛʃ/, architect /ˈɑɹ.kɪˌtɛkt/, analyst /ˈænəlɪst/, introduce /ˌɪntɹəˈdus/
/r/ 和 /l/ 的區別: virtual /ˈvɜɹt͡ʃuəl/, centrally /ˈsɛnt.ɹə.li/

如何用這段影片練習

先完整聽一遍影片，不要開口，記下不認識的單詞。
先用 0.75 倍速逐句跟讀，熟練之後再回到正常速度。
錄下自己的聲音並與原聲對比，特別注意 catalogue, domain, marketplace 這類單詞。

什麼是跟讀法？

跟讀法 (Shadowing) 是一種有科學依據的語言學習技巧，最初開發用於專業口譯員的培訓，並由多語言者Alexander Arguelles博士普及。這個方法簡單而強大：您在聽英語母語原聲的同時立即大聲重複——就像是一個延遲1-2秒緊跟說話者的影子。與被動聽力或語法練習不同，跟讀法強迫您的大腦和口腔肌肉同時處理並模模擬實的講話模式。研究表明它能顯著提高發音準確性，語調，節奏，連讀，聽力理解和口語流利度——使其成為雅思口語備考和真實英語交流最有效的方法之一。

影子跟讀法: 閱讀完整分步指南 →

跟讀練習: What Is a Data Mesh? Databricks Unity Catalog Explained - 通過影片學習英語口語

選擇練習模式

下載應用

同頻道

How Modern Data Teams Work (Engineer, Analyst, Scientist, Architect, ML)

Learn Databricks for FREE (Step-by-Step Guide)

推薦影片

Database Sharding! Designing Data-Intensive Applications chapter 6

Why Abu Dhabi Wants to Export Energy AI

How Anthropic uses Claude in Product Management

The Dropdown Slicer We Waited 10 Years For (Power BI)

Running AI Agents in Secure Sandboxes with E2B & Docker MCP | Docker’s AI Guide to the Galaxy

What is Event Driven Architecture (EDA)?

what is a database index?

Database vs Data Warehouse vs Data Lake | Explaining the Differences

THIS Is Where the Internet Lives

Agentic AI Frameworks Explained: Workflows, Multi-Agent, & Production

Sec Growth DataScience staff meeting Sep 14 2022

Understand Microsoft Entra ID | AZ-104 | Episode 2