Data Types

All supported data types in Strata semantic models.

Overview

Strata supports 8 data types for dimensions and measures. Choose the type that matches your database column type.

Supported Types

string

Text values. Use for names, descriptions, codes, etc.

- type: dimension
  name: Customer Name
  data_type: string
  expression:
    sql: customer_name

Use cases:

• Names (customer, product, etc.)
• Descriptions
• Codes (SKU, order number as text)
• Categories
• Status values

integer

Whole numbers. Use for counts, IDs, quantities.

- type: dimension
  name: Order ID
  data_type: integer
  expression:
    primary_key: true
    sql: order_id

- type: measure
  name: Order Count
  data_type: integer
  expression:
    sql: count(*)

Use cases:

• IDs (order, customer, product)
• Counts
• Quantities
• Years

Range: -2,147,483,648 to 2,147,483,647

bigint

Large whole numbers. Use for very large integers.

- type: dimension
  name: Customer ID
  data_type: bigint
  expression:
    sql: customer_id

Use cases:

• Very large IDs
• Timestamps (Unix epoch)
• Large counts

Range: -9,223,372,036,854,775,808 to 9,223,372,036,854,775,807

decimal

Floating-point numbers. Use for monetary values, percentages, measurements.

- type: measure
  name: Total Revenue
  data_type: decimal
  format: currency:2
  expression:
    sql: sum(amount)

Use cases:

• Money/currency
• Percentages
• Measurements (weight, distance, etc.)
• Ratios
• Averages

Precision: Database-dependent (typically 38 digits)

date

Date values (without time). Use for date dimensions.

- type: dimension
  name: Order Date
  data_type: date
  grains:
    - day
    - week
    - month
    - quarter
    - year
  expression:
    sql: order_date

Use cases:

• Date dimensions
• Order dates
• Birth dates
• Any date without time

Format: YYYY-MM-DD

Grains: day, week, month, quarter, year

date_time

Timestamp values (with time). Use for precise timestamps.

- type: dimension
  name: Created At
  data_type: date_time
  grains:
    - raw
    - hour
    - day
    - week
    - month
  expression:
    sql: created_at

Use cases:

• Timestamps
• Created/updated dates
• Event times
• Any date with time

Format: YYYY-MM-DD HH:MM:SS

Grains: raw, second, minute, hour, day, week, month, quarter, year

boolean

True/false values.

- type: dimension
  name: Is Active
  data_type: boolean
  expression:
    sql: is_active

Use cases:

• Flags (is_active, is_deleted, etc.)
• Binary states
• Yes/no fields

Values: true, false, NULL

binary

Binary data. Rarely used in analytics.

- type: dimension
  name: Image Data
  data_type: binary
  expression:
    sql: image_blob

Use cases:

• Binary files
• Images (prefer storing URLs as string)
• Encrypted data

Note: Rarely used. Prefer storing references (URLs, file paths) as strings.

Date and DateTime Grains

For date and date_time types, specify supported granularities:

grains:
  - day
  - week
  - month
  - quarter
  - year

Available grains:

• raw - Original value (date_time only)
• second - Second-level granularity (date_time only)
• minute - Minute-level granularity (date_time only)
• hour - Hour-level granularity (date_time only)
• day - Day-level granularity
• week - Week-level granularity
• month - Month-level granularity
• quarter - Quarter-level granularity
• year - Year-level granularity

Example:

- type: dimension
  name: Order Date
  data_type: date
  grains:
    - day
    - week
    - month
    - quarter
    - year
  expression:
    sql: order_date

Type Selection Guide

Database Type	Strata Type	Notes
VARCHAR, TEXT	string	Text values
INT, INTEGER	integer	Whole numbers
BIGINT	bigint	Large integers
DECIMAL, NUMERIC, FLOAT, DOUBLE	decimal	Floating-point
DATE	date	Date without time
TIMESTAMP, DATETIME	date_time	Date with time
BOOLEAN, BOOL	boolean	True/false
BLOB, BINARY	binary	Binary data (rare)

Best Practices

1. Match database types - Use the corresponding Strata type
2. Use decimal for money - Never use float for currency
3. Set grains for dates - Enable flexible date grouping
4. Use integer for IDs - Unless you need bigint
5. Avoid binary - Prefer storing references as strings

Next Steps

• Learn about dimensions
• Learn about measures
• Understand field format
• Read about expressions