Domain 1.3 of the DA0-002 exam asks you to identify and categorise the fundamental building blocks of data. Understanding data types, data structures, and metadata is the difference between a messy spreadsheet you cannot search and a well-organised database you can query in seconds. This concept matters because every analysis, every report, and every business decision starts with knowing what kind of data you have and how it is arranged.
Jump to a section
A simple way to picture Data Types, Data Structures, and Metadata
A municipal urban planner is organising the city's building permits. She has a giant filing cabinet with three distinct sections. The first section contains blueprints. Each blueprint is a specific *type* of document: a house blueprint is a different size and format from a skyscraper blueprint, and both are different from a paved-street blueprint. She cannot file a house blueprint in the skyscraper slot, because the format is wrong. That is your data type: the strict format and kind of data you are allowed to store (like a number versus a date versus a sentence).
The second section of the filing cabinet is the *structure*. It contains folders. One folder holds all permits for a single housing estate: each permit is a row, and each row has columns for address, owner name, permit date, and permit cost. That folder is a table: a rigid, organised arrangement of rows and columns. Another folder holds a map of the city, with pins on it for each building. The map, with its pins connected by lines, is a graph: a structure for showing relationships between things. The planner chooses the folder structure based on how she needs to search and relate the data.
The third section is a set of sticky notes attached to each file. The sticky notes say things like: "Filed on 3rd April 2025 by Park Services", "Contains private data: do not share", and "Last viewed by Building Inspector Kim". These notes are metadata: they are data *about* the data. They don't describe the house itself, but they describe the file that describes the house. Without the sticky notes, the planner would not know who filed the permit, or when it was last checked, or that it contains restricted information. The planner, the data, and the sticky notes together make the city run smoothly; without them, permits would be lost, formats would be jumbled, and nobody would trust the records.
Data types, data structures, and metadata are the three pillars of organising information. They are not the same thing, but they work together. Think of data types as the rules for what goes into a single cell of a spreadsheet (text, number, date). Data structures are the containers that hold those cells (tables, lists, trees). Metadata is the documentation that describes the containers and their contents (column names, creation dates, permissions).
Let us start with data types. A data type is a classification that tells a computer or database what kind of value you are storing and what operations you can perform on it. The most common data types you will see in the DA0-002 exam include:
String: a sequence of characters, like "John Smith" or "A123B". Strings hold text. You can join them together (concatenation) but you cannot multiply them.
Integer: a whole number without a decimal point, like 42 or -5. Integers are used for counting and indexing.
Float: a number with a decimal point, like 3.14 or -0.001. Floats are used for measurements and calculations where precision matters.
Boolean: a logical value that is either True or False. Booleans are used for yes/no questions, flags, and condition checks.
Date/Time: a specific point in time, like "2025-04-14" or "14:30:00". Dates and times have their own rules for arithmetic (you can subtract one date from another to get days, but you cannot add two dates).
Character: a single letter, digit, or symbol. It is like a string but only one character long. You will rarely see it tested on its own, but it exists.
Why do data types matter? Because if you try to add a string like "apple" to an integer like 5, the computer will either throw an error or give you a nonsensical result. Data types are the computer's way of preventing you from making category mistakes. When you design a database table, you must assign a data type to each column. The DA0-002 exam loves to test whether you know which data type fits a given piece of information.
Now on to data structures. A data structure is an organised way of storing and accessing data. It defines the relationship between data elements and the operations you can perform on them. The DA0-002 exam focuses heavily on three data structures:
Tables (also called relations or spreadsheets): data organised into rows and columns. Each row is a record, and each column is a field. Tables are the core of relational databases. They are good when you have structured data that fits into fixed categories, like customer names, addresses, and order amounts.
Arrays/Lists: ordered collections of items, where each item has a position (index). A list of student test scores [85, 92, 78] is an array. Lists are simple and fast to read from start to finish, but hard to search if you do not know the position.
Trees: hierarchical structures where each item (node) is connected to one parent and possibly many children. A company org chart is a tree. Trees are excellent for representing categories, taxonomies, and file systems. The DA0-002 exam often uses trees to test your understanding of parent-child relationships.
Data structures are chosen based on how you plan to use the data. If you need to look up a value by a unique key (like a customer ID), a table with an index is best. If you need to store a sorted list of names and iterate through them, an array works. If you need to show a product category with its sub-categories, you use a tree.
Metadata is data that describes other data. It is the context that makes raw data understandable. There are three common types of metadata you will see on the exam:
Descriptive metadata: describes the content for discovery and identification. For example, an email's subject line, author name, and date sent.
Structural metadata: describes how the data is organised. For example, the fact that a book has chapters, and that each chapter contains sections, is structural metadata.
Administrative metadata: describes the management of the data, such as who created it, when it was last modified, and who can access it. For example, a file's permissions and creation timestamp.
Metadata is essential for data governance. Without metadata, you might have a column named "date" with no way of knowing if it means the date the record was created, the date the sale happened, or the date the data was last updated. Metadata answers those questions. On the DA0-002 exam, you will be asked to identify which piece of information is metadata versus which is actual data. For example, the number of rows in a table is metadata; the value within a row is data.
Identify the Content Type
You start by looking at a piece of information and deciding what category it belongs to. For example, you see '42'. Is it text, a number, or something else? This is your first step because data types are the foundation. If you misidentify the type, every subsequent step—storing, filtering, calculating—will be wrong. In the DA0-002 exam, the first skill tested is recognising whether a value is a string, integer, float, boolean, date, time, or character.
Select the Correct Data Type for Storage
Once you know the content is a number like 3.14, you must choose the exact data type: float, not integer. This step matters because choosing integer would truncate the decimal part, destroying precision. If you have a phone number like '555-1234', you choose string, not integer, to preserve the dash and leading zeros. This decision directly impacts data integrity—storing data in the wrong type corrupts it.
Choose an Appropriate Data Structure
After you know your data types, you decide how to organise the collection of data. If you have a list of employee names and their salaries, a table (two columns, each with its own data type) is the right structure. If you need to model a product category hierarchy (Electronics > Computers > Laptops), you choose a tree. This step is critical because a poor structure makes queries slow and data hard to maintain.
Normalise the Data Structure (If Necessary)
If you chose a table, you now split it into related tables to remove redundancy. For instance, instead of storing customer address in every order row, you create a separate Customers table and link it via a customer ID. This step improves efficiency and consistency. The DA0-002 exam tests whether you understand why normalisation (splitting data) is important for data quality.
Create and Attach Metadata
Finally, you document the data. You write a data dictionary that lists each column, its data type, a description, and any rules (like 'not null'). You also add administrative metadata: who created the table, when it was last modified, and who can access it. This step ensures that anyone else—or your future self—can understand and trust the data. Metadata is what turns a raw dataset into a shared business asset.
Sarah is a junior data analyst at a mid-sized company that sells office furniture online. She has been given a raw CSV file containing customer orders from the last quarter. Her first task is to prepare this data for analysis. Here is what she actually does with data types, data structures, and metadata.
First, Sarah opens the CSV file in a spreadsheet tool. She looks at each column and recognises that the data types are mixed. The column "CustomerID" should be a string because it contains letters and numbers (like "CUST-3452"). The column "OrderAmount" should be a float because it contains values like 149.99. But the spreadsheet she received has "OrderAmount" stored as text, with a currency symbol and commas ("$1,349.00"). That is a problem. If she tries to calculate the average order amount, the function will fail or return zero because the computer sees text, not numbers. Sarah must convert the data type of that column to a float before she can analyse it. She also spots that the "OrderDate" column is in a mix of formats: some rows say "2025-03-14", some say "14/03/2025", and a few are blank. She standardises all dates to a single format (YYYY-MM-DD) and assigns the correct date data type. This step takes her 45 minutes, but it saves days of debugging later.
Next, Sarah needs to structure the data. The raw CSV is one giant flat table with 15 columns. That is the simplest data structure: a table. But Sarah needs to join this data with another table containing customer information, and with a third table containing product details. She realises that trying to do everything in one flat table would cause massive duplication. For example, if a customer orders three items, their name and address are repeated three times. So she decides to use a relational data structure: a database with separate tables for Customers, Orders, and Products. She creates the tables, defines primary keys (a unique identifier for each row, like CustomerID), and foreign keys (a column in one table that references a primary key in another table, like CustomerID in the Orders table linking back to the Customers table). This is a classic normalisation step—removing redundancy by splitting data across related tables. The data structure she chooses (relational tables with keys) directly impacts how fast she can query the data and how consistent it stays.
Finally, Sarah adds metadata. For each table, she writes a data dictionary—a document that lists every column, its data type, a description, and an example value. For instance, she writes: "Column: CustomerEmail. Data type: String. Description: The primary email address of the customer. Example: john.doe@example.com." She also adds administrative metadata: timestamps for when each row was last updated, and a flag for which rows have been validated. Without this metadata, anyone else who looks at Sarah's tables would have to guess what each column means. Metadata turns a pile of raw numbers and text into a shared, understandable resource. When her manager asks, "What does the 'status' column mean? Are there only two possible values?", Sarah can point to the metadata and answer immediately.
Sarah's work does not stop there. She also uses metadata to filter the data she is allowed to see. The company has a rule: any order where the customer has opted out of marketing cannot have their email or phone number used in analysis. That rule is stored as metadata. Sarah's query tool checks the metadata before showing her the data, ensuring she never accidentally violates privacy rules. This is a real-world example of data governance metadata at work.
The DA0-002 exam will test you on Domain 1.3 with a mix of multiple-choice and multiple-select questions. The exam loves to throw in traps where they give you a real-world scenario and ask you to identify the data type, data structure, or metadata element being described. Here is exactly what you need to know to get those points.
First, know the seven core data types by heart: string, integer, float, boolean, date, time, and character. The exam almost always includes a question like: "A field stores a value of 'True' or 'False'. Which data type is most appropriate?" The answer is boolean. Another common trap: they give you a field storing "2025-04-14" and ask if it is a string or a date. The answer is date if you intend to perform date arithmetic, but if you are just displaying it as text, it could be a string. The exam will specify the use case, so read carefully. They will also test float vs integer: if you need to store a price like $9.99, you need a float (or decimal), not an integer.
Common data type traps:
They ask: "Which data type would you use for a customer phone number?" Beginners often say integer. Wrong. Phone numbers are strings because they may contain dashes, parentheses, or leading zeros. You never add or subtract phone numbers.
They ask: "Which data type for a primary key?" If the key is auto-generated as a number, integer is correct. But if the key is an employee code like "E00123", it is a string.
They ask: "Which data type for a discount percentage like 0.15?" Float is correct, not integer.
Second, data structures appear in questions about how data is organised. The exam focuses on tables (relations), arrays/lists, and trees. A typical question: "A company's organisational chart shows a CEO at the top, managers below, and employees under each manager. Which data structure best represents this?" The answer is a tree. Another question: "You have a list of customers that you need to iterate through in the order they were added. Which data structure is simplest?" Answer: an array or list. They may also ask about the difference between structured data (which fits neatly into tables) and unstructured data (like images or free text, which does not). Know that a database table is a structure for storing structured data, while a data lake is a place for storing unstructured data.
Third, metadata is a favourite topic. The exam will ask you to distinguish between data and metadata. For example: "A spreadsheet contains sales figures for each month. The column headers 'Month' and 'Sales' are an example of what?" The answer is structural metadata. Another classic: "The file creation date and the last modified timestamp are examples of which type of metadata?" Administrative. They will also test descriptive metadata: "An author's name and a document's title are examples of which metadata type?" Descriptive. The trap here is that beginners confuse metadata with the data itself. For instance, they might say that a column of numbers is metadata. It is not—the numbers are the data. Metadata is the information *about* that column, like its name, its data type, and its constraint rules.
Exam definitions to memorise:
Data type: The category of a value (string, integer, float, boolean, date, time, character).
Data structure: The organisational format for storing and accessing data (table, array/list, tree).
Metadata: Data about data, including descriptive, structural, and administrative types.
Structured data: Data that fits into a fixed schema of rows and columns.
Unstructured data: Data without a predefined format, such as images, videos, and raw text files.
The DA0-002 exam also tests your understanding of why these concepts matter. They might present a scenario: "A manager complains that reports take too long to generate because the database stores all customer information in one huge table with repeated data. What is the most likely data structure issue?" The answer is lack of normalisation or improper use of a data structure (they should use separate tables with relations). Another scenario: "A data analyst cannot tell whether a 'DATE' column contains the order date or the shipping date. What is missing?" The answer is metadata (a description of the column).
Finally, expect a question about the purpose of metadata in data governance. The correct answer is that metadata supports data lineage (tracking where data came from), data quality (knowing when data was last updated), and data security (knowing who can access it). They will never ask you to memorise obscure metadata standards; they just want you to understand the concept and its role.
Data types enforce what kind of value can be stored in a field, preventing category errors like adding text to a number.
The six essential data types for DA0-002 are string, integer, float, boolean, date, and time — each with specific use cases.
Data structures determine how data is organised and accessed; tables are for structured data, arrays for ordered lists, and trees for hierarchies.
Metadata is data about data, and it comes in three flavours: descriptive, structural, and administrative.
Structured data fits neatly into rows and columns, while unstructured data (like images or PDFs) does not have a predefined format.
Choosing the wrong data type or data structure leads to inefficient storage, incorrect calculations, and difficulty querying the data.
A data dictionary is a practical example of metadata that documents column names, data types, descriptions, and constraints.
On the exam, always ask yourself: 'Is this piece of information describing the actual value, or is it describing the record itself?' — that will separate data from metadata.
These come up on the exam all the time. Here's how to tell them apart.
String
Stores text, symbols, or digits as characters.
Cannot be used for arithmetic like addition or subtraction.
Preserves leading zeros (e.g., '00123' stays as 00123).
Integer
Stores whole numbers without a decimal point.
Can be used for mathematical operations like adding and multiplying.
Drops leading zeros (e.g., 00123 becomes 123).
Float
Stores numbers with a decimal point (e.g., 3.14).
Used for precise measurements like prices or scientific data.
Takes more storage space because it holds fractional parts.
Integer
Stores whole numbers only (e.g., 3).
Used for counts, IDs, and quantities that are always whole.
Takes less storage space and is faster for calculations.
Table (Relation)
Data is arranged in rows and columns with no inherent hierarchy.
Relationships between rows are established through foreign keys.
Best for many-to-many relationships and simple lookups by key.
Tree
Data is arranged in nodes with a parent-child hierarchy.
Each node has exactly one parent (except the root).
Best for representing organisational charts, file systems, and taxonomies.
Administrative Metadata
Includes information about management: who created the file, when it was last modified, and access permissions.
Used for security, auditing, and data governance.
Answers questions like 'Who owns this data?' and 'When was it updated?'
Descriptive Metadata
Includes information for discovery and identification: title, author, keywords, subject.
Used for search and retrieval (like finding a document by its title).
Answers questions like 'What is this data about?' and 'Who wrote it?'
Structured Data
Has a fixed schema: data fits into rows and columns with defined data types.
Examples include spreadsheets, SQL databases, and CSV files.
Easy to query, filter, and analyse using standard tools.
Unstructured Data
Has no predefined format or schema.
Examples include images, videos, PDFs, and raw text files.
Requires special processing (like natural language processing or image recognition) to extract meaning.
Boolean
Holds only two values: True or False.
Used for logical operations and binary conditions (e.g., 'isActive').
Extremely storage-efficient (1 bit per value in many systems).
String
Holds any sequence of characters, including 'True' and 'False' as text.
Used for names, addresses, codes, and free text.
Requires more storage and cannot be used directly in logical expressions.
Mistake
Metadata is just the column headers in a spreadsheet.
Correct
Column headers are part of metadata, but metadata is much broader. It includes structural metadata (headers, table relationships), administrative metadata (creation date, owner, permissions), and descriptive metadata (tags, descriptions, keywords). An entire data dictionary is metadata, not just a single row of headers.
Beginners see metadata only as labels because spreadsheets commonly have a header row. They do not realise that metadata operates at multiple levels (file, table, column, row) and that databases store metadata in separate system tables.
Mistake
A 'string' data type and a 'character' data type are the same thing.
Correct
A character holds exactly one character (like 'A' or '5'), while a string can hold any number of characters from zero to thousands. In many databases, CHAR and VARCHAR are distinct data types. The DA0-002 exam treats them as separate types, and you will be tested on knowing which one to use for a single letter versus a word or sentence.
Because in everyday language we say 'string of characters', beginners blur the distinction. They do not know that a character field has a fixed length of 1 and is more efficient for single-letter codes like gender (M/F) or a flag (Y/N).
Mistake
A Boolean data type can store 'Yes' or 'No' as text.
Correct
A Boolean data type holds only two values: True or False. It is a logical type, not a text type. If you need to store 'Yes' or 'No' as words, you must use a string data type, not Boolean. Storing Boolean as text wastes space and loses the ability to use logical operations (AND, OR, NOT).
Beginners confuse the concept of a binary choice (yes/no) with the actual data type implementation. Many user interfaces display checkboxes as 'Yes/No', so people assume a database can store them as text. They do not realise that True/False is a special type that takes only 1 bit of storage.
Mistake
A tree data structure and a table data structure are interchangeable if you use the right query.
Correct
They are fundamentally different. A table is a flat set of rows and columns with no inherent hierarchy (though relations can be created through foreign keys). A tree is a hierarchical structure where each node has one parent and zero or more children. You cannot represent a true hierarchy (like an organisation chart) efficiently in a single table without repeating data or using complex recursive queries. The data structure you choose determines what operations are fast and natural.
Beginners think all data can be forced into a spreadsheet because spreadsheets are familiar. They do not understand that tree operations (like getting all descendants of a node) are simple in a tree but require complex self-joins in a table.
Mistake
If I have a file named 'sales_report_2025.csv', the file name is part of the data.
Correct
The file name is metadata, not data. The data is the actual sales figures, customer names, and dates inside the CSV file. The file name describes the data (it tells you the file contains a sales report from 2025), but it is not itself a sales record. Metadata includes file names, file sizes, creation dates, and permissions.
Beginners think of everything in a file system as 'data'. They do not distinguish between the content and the container. The exam tests this exact boundary: they ask 'Which is metadata?' and list both the file creation date and a sales value. You must pick the creation date.
Reveal each answer, then mark whether you got it right. Score 60%+ to unlock the next chapter.
A character holds exactly one character (like 'A' or '5'), while a string holds a sequence of zero or more characters (like 'Hello' or 'A123'). Use character for single-letter codes (M/F) and string for words, sentences, or codes with multiple characters.
You can, but it is bad practice. If you store a date as a string, you cannot perform date arithmetic (like calculating days between two dates) and you risk inconsistent formats (e.g., '03/04/2025' vs '4 March 2025'). The exam expects you to use the date data type when the field will be used for date operations.
Structural metadata describes how data is organised. For example, the fact that a database table has columns called 'CustomerName', 'OrderDate', and 'TotalAmount', and that the table is related to another table through a foreign key, is structural metadata. It tells you the schema or layout of the data.
No, a primary key is a constraint, not a data type. The data type of a primary key can be an integer (if it is an auto-incrementing number) or a string (if it is a natural key like an employee code). The exam tests that you know the difference between data types (the kind of value) and database constraints (rules applied to columns).
A family tree is a perfect example of a tree data structure. Each person (node) has one parent (except the root) and can have multiple children. A table would be inefficient because you would need to repeatedly query to find all descendants. Trees naturally represent hierarchical relationships.
Yes, you should memorise the seven common data types: string, integer, float, boolean, date, time, and character. The exam will give you scenarios and ask you to pick the correct type. Also memorise that names (strings) and IDs with letters (strings) are the most common traps.
You've finished Data Types, Data Structures, and Metadata. Continue through the DA0-002 study guide to build a complete picture of the exam.
Done with this chapter?