ILLEGAL_TYPE_OF_ARGUMENT ClickHouse error

This error usually means you're passing the wrong type to a function or operator. ClickHouse is strict about types, and many functions require exact matches.

The ILLEGAL_TYPE_OF_ARGUMENT error in ClickHouse (and Tinybird) happens when a function receives an argument of the wrong type. For example, trying to sum a String, or using toDate on an Int without casting.

This is one of the most common issues users hit when building Pipes or querying Tinybird Data Sources.

What causes this error

You'll typically see it when:

  • Using a function with the wrong input type (e.g., sum(String)).
  • Implicit casting is not supported (ClickHouse won't auto-convert).
  • You're mixing types in expressions (String + Int, DateTime - String, etc).
  • Aggregations or filters using values that don't match column types.

Example errors

Fails: wrong type in sum()
SELECT sum(url) FROM events
Fails: toDate() on Int
SELECT toDate(1234567890)
Fails: substring() on Int
SELECT substring(123, 1, 5)
Fails: DateTime minus String
SELECT now() - '5'

How to fix it

Use explicit casts and make sure argument types match the function expectations:

Fix: cast String to Float64
SELECT sum(toFloat64(url)) FROM events
Fix: convert Int to DateTime first
SELECT toDate(toDateTime(1234567890))
Fix: subtract seconds using interval
SELECT now() - INTERVAL 5 SECOND

Common casting functions

ClickHouse doesn't automatically cast strings to numbers. Use these functions for explicit type conversion:

FunctionPurposeExample
toInt32(), toInt64()Convert to integertoInt32('123')
toFloat32(), toFloat64()Convert to floattoFloat64('123.45')
toString()Convert to stringtoString(123)
toDate(), toDateTime()Convert to date/datetimetoDateTime('2023-01-01 12:00:00')
parseDateTimeBestEffort()Smart datetime parsingparseDateTimeBestEffort('2023/01/01 12:00:00')
toInt32OrZero(), toFloat64OrZero()Safe conversion (returns 0 on failure)toInt32OrZero('abc') → 0

Safe conversion with try functions

Use the try family of functions when you're unsure about data quality:

Safe conversion examples
SELECT toInt32OrZero('123') as safe_int,      -- Returns 123
       toInt32OrZero('abc') as failed_int,    -- Returns 0
       toFloat64OrZero('123.45') as safe_float -- Returns 123.45

Proper datetime handling

For datetime operations, use proper functions:

Correct datetime examples
SELECT toDateTime('2023-01-01 12:00:00') as parsed_datetime,
       parseDateTimeBestEffort('2023/01/01 12:00:00') as smart_parsed,
       now() - INTERVAL 5 SECOND as five_seconds_ago

Tinybird-specific notes

In Pipes, the schema inference might auto-assign wrong types if input data is inconsistent. You can:

  • Use toTypeName() to debug columns and check what's being inferred.
  • Add a CAST() early in the pipe to normalize types.
  • Validate incoming types in Data Sources via Schema Hints.

See also