Skip to main content

Transformation: Parse Date

Convert date values from any format into a standardized datetime format, including Unix epoch timestamps in seconds or milliseconds.

Updated this week

The Parse Date transformation converts date values in your dataset from one format to another, producing a clean, standardized output. This is useful any time your source data contains dates in an inconsistent or system-specific format, such as Unix epoch timestamps from an API, US-style date strings from a CRM export, or month-name formats from a reporting tool.

Parse Date replaces the selected column in place rather than creating a new one, keeping your dataset organized.

How to Add a Parse Date Transformation

In your pipeline, click Select transformation and choose Parse Date from the list. The transformation form will appear with the following fields:

  • Name: Give your transformation a descriptive name, such as "Parse Created At" or "Convert Epoch to Date."

  • Column: Select the single column containing the date values you want to convert.

  • Input Format: Enter the format code that matches how the dates are currently stored in your data. See the examples below for common formats.

  • Output Format: Enter the format code for how you want the dates to appear after transformation. Defaults to %Y-%m-%d %H:%M:%S.

  • Output Timezone: Select the timezone for the formatted output. Defaults to UTC.

Once configured, click Save Transformation to apply.

The Parse Date transformation form showing Name, Column, Input Format, Output Format, and Output Timezone fields

Common Business Use Cases

Converting Unix Epoch Timestamps (API and Event Data)

Many APIs and data platforms store timestamps as Unix epoch values -- the number of seconds or milliseconds elapsed since January 1, 1970. These are not human-readable without conversion.

Scenario

Input Format

Output Format

Example Input

Example Output

Epoch in seconds (e.g., Salesforce, Segment)

%s

%Y-%m-%d %H:%M:%S

1709596800

2024-03-05 00:00:00

Epoch in milliseconds (e.g., Mixpanel, Kafka)

%Q

%Y-%m-%d %H:%M:%S

1709596800000

2024-03-05 00:00:00

Note: Use %s for epoch timestamps in seconds and %Q for timestamps in milliseconds. If your values are 13 digits, use %Q. If they are 10 digits, use %s.

Reformatting US-Style Date Strings (CRM and Spreadsheet Exports)

Data exported from tools like Salesforce, HubSpot, or Excel often uses a US date format (month/day/year). Parse Date lets you standardize these to ISO format for downstream processing.

Scenario

Input Format

Output Format

Example Input

Example Output

US date format

%m/%d/%Y

%Y-%m-%d

03/05/2024

2024-03-05

US date with time

%m/%d/%Y %H:%M:%S

%Y-%m-%d %H:%M:%S

03/05/2024 14:30:00

2024-03-05 14:30:00

Converting Long-Form Date Strings (Reports and Documents)

Some reporting tools or document exports write out the full month name. Parse Date handles these as well.

Scenario

Input Format

Output Format

Example Input

Example Output

Full month name

%B %d, %Y

%Y-%m-%d

March 05, 2024

2024-03-05

Abbreviated month

%b %d, %Y

%Y-%m-%d

Mar 05, 2024

2024-03-05

Extracting Date Only from a Datetime Value

If your source data includes a full timestamp but you only need the date portion for grouping or reporting purposes, you can use Parse Date to strip the time.

Scenario

Input Format

Output Format

Example Input

Example Output

Datetime to date only

%Y-%m-%d %H:%M:%S

%Y-%m-%d

2024-03-05 09:45:00

2024-03-05

Format Code Reference

Parse Date uses strftime format codes. The most common codes are:

  • %Y -- Four-digit year (e.g., 2024)

  • %m -- Two-digit month (e.g., 03)

  • %d -- Two-digit day (e.g., 05)

  • %H -- Hour in 24-hour format (e.g., 14)

  • %M -- Minutes (e.g., 30)

  • %S -- Seconds (e.g., 00)

  • %B -- Full month name (e.g., March)

  • %b -- Abbreviated month name (e.g., Mar)

  • %s -- Unix epoch in seconds

  • %Q -- Unix epoch in milliseconds

Note: If a value in the selected column is null or cannot be parsed using the specified input format, Parse Date will convert that value to null rather than returning an error.

Did this answer your question?