TEST_EXECUTION_RESULTS (Single-Value Tests)

Overview

RESULTS.TEST_EXECUTION_RESULTS stores the row-level output for single-value tests — tests that compare one scalar value from the test dataset against one scalar value from the control dataset. Examples include row count comparisons, aggregate checks (SUM, AVG, MIN, MAX), and any test where the data set column type is set to produce a single-row result.

Each row in this table represents one comparison within a test execution. Most single-value tests produce exactly one row per execution, but some (like profile-based tests with a coverage flag) may produce multiple rows.

Schema

Column	Type	Nullable	Description
TEST_EXECUTION_ID	NUMBER	No	Links to VW_DATASET_TEST_EXECUTIONS — identifies which execution produced this result
TEST_ID	NUMBER	No	The test ID
TEST_VERSION_ID	NUMBER	No	The test version ID
ROW_INDEX	NUMBER	No	Row position within this execution (usually 1 for single-value tests)
COVERAGE_FLAG	NUMBER	No	Coverage type: 3 = Matched (both source and target returned data)
OVERFLOW_FLAG	NUMBER	No	0 = normal, 1 = value truncated due to size
COMBINED_KEY	VARCHAR	Yes	Combined key value (usually NULL for single-value tests)
NUMERIC_SOURCE_VALUE	DECIMAL(27,8)	Yes	Numeric value from the test (source) dataset
NUMERIC_TARGET_VALUE	DECIMAL(27,8)	Yes	Numeric value from the control (target) dataset
ACTUAL_DIFFERENCE	DECIMAL(27,8)	Yes	Source minus target
ABSOLUTE_DIFFERENCE	DECIMAL(27,8)	Yes	Absolute value of the difference
PERCENT_DIFFERENCE	DECIMAL(18,9)	Yes	Percentage difference (as a decimal, e.g., 0.05 = 5%)
EVALUATION_RESULT_ID	NUMBER	Yes	Result code: 24 = Passed, 25 = Failed, 26 = Error
EVALUATION_RESULT_NAME	VARCHAR(100)	Yes	"PASSED", "FAILED", or "ERROR"
STRING_SOURCE_VALUE	VARCHAR	Yes	String value from the source (when comparing strings)
STRING_TARGET_VALUE	VARCHAR	Yes	String value from the target
DATE_SOURCE_VALUE	VARCHAR	Yes	Date value from the source (stored as string)
DATE_TARGET_VALUE	VARCHAR	Yes	Date value from the target
SF_DATE_CREATED	TIMESTAMP	No	When this row was loaded into Snowflake
SF_IMPORT_FILENAME	VARCHAR(500)	Yes	Staging file that delivered this row

How Values Are Stored

The table uses three pairs of columns depending on the data type being compared:

Comparison Type	Source Column	Target Column	Difference Columns
Numeric	NUMERIC_SOURCE_VALUE	NUMERIC_TARGET_VALUE	ACTUAL_DIFFERENCE, ABSOLUTE_DIFFERENCE, PERCENT_DIFFERENCE
String	STRING_SOURCE_VALUE	STRING_TARGET_VALUE	(not applicable — pass/fail only)
Date	DATE_SOURCE_VALUE	DATE_TARGET_VALUE	(not applicable — pass/fail only)

For numeric comparisons, the difference columns are pre-calculated. For string and date comparisons, the result is strictly pass/fail based on equality.

Relationships

RESULTS.TEST_EXECUTION_RESULTS
    │
    ├── TEST_EXECUTION_ID → REPOSITORY.VW_DATASET_TEST_EXECUTIONS.TEST_EXECUTION_ID
    ├── TEST_ID → REPOSITORY.VW_DATASET_TESTS.TEST_ID
    └── TEST_VERSION_ID → REPOSITORY.VW_DATASET_TESTS.TEST_VERSION_ID

Example Queries

Get Source/Target/Difference for a Specific Execution

SELECT
    TEST_EXECUTION_ID,
    ROW_INDEX,
    NUMERIC_SOURCE_VALUE,
    NUMERIC_TARGET_VALUE,
    ACTUAL_DIFFERENCE,
    PERCENT_DIFFERENCE,
    EVALUATION_RESULT_NAME
FROM RESULTS.TEST_EXECUTION_RESULTS
WHERE TEST_EXECUTION_ID = 3913133;

Join to Test Executions for Context

SELECT
    te.TEST_NAME,
    te.PROJECT_NAME,
    te.DATE_COMPLETED,
    r.NUMERIC_SOURCE_VALUE AS SOURCE_VALUE,
    r.NUMERIC_TARGET_VALUE AS TARGET_VALUE,
    r.PERCENT_DIFFERENCE,
    r.EVALUATION_RESULT_NAME
FROM REPOSITORY.VW_DATASET_TEST_EXECUTIONS te
JOIN RESULTS.TEST_EXECUTION_RESULTS r
    ON r.TEST_EXECUTION_ID = te.TEST_EXECUTION_ID
WHERE te.LATEST_TEST_EXECUTION_INDEX = 1
    AND te.TEST_NAME = 'Row Count - DIM_CUSTOMER';

Trend Analysis: Track Values Over Time

SELECT
    te.TEST_NAME,
    te.DATE_COMPLETED,
    r.NUMERIC_SOURCE_VALUE,
    r.NUMERIC_TARGET_VALUE,
    r.PERCENT_DIFFERENCE,
    r.EVALUATION_RESULT_NAME
FROM REPOSITORY.VW_DATASET_TEST_EXECUTIONS te
JOIN RESULTS.TEST_EXECUTION_RESULTS r
    ON r.TEST_EXECUTION_ID = te.TEST_EXECUTION_ID
WHERE te.TEST_ID = 58874
ORDER BY te.DATE_COMPLETED;

Find All Failed Single-Value Tests with Their Differences

SELECT
    te.TEST_NAME,
    te.TEST_PRIMARY_METADATA_LINK_OBJECT,
    te.DATE_COMPLETED,
    r.NUMERIC_SOURCE_VALUE,
    r.NUMERIC_TARGET_VALUE,
    r.ACTUAL_DIFFERENCE,
    ROUND(r.PERCENT_DIFFERENCE * 100, 2) AS PERCENT_DIFF_DISPLAY
FROM REPOSITORY.VW_DATASET_TEST_EXECUTIONS te
JOIN RESULTS.TEST_EXECUTION_RESULTS r
    ON r.TEST_EXECUTION_ID = te.TEST_EXECUTION_ID
WHERE te.LATEST_TEST_EXECUTION_INDEX = 1
    AND r.EVALUATION_RESULT_NAME = 'FAILED'
ORDER BY ABS(r.PERCENT_DIFFERENCE) DESC;

Databricks Notes

Table structure and column names are identical between Snowflake and Databricks
ROUND() function syntax is the same
The SF_DATE_CREATED and SF_IMPORT_FILENAME columns are Snowflake-specific metadata; in Databricks these may be named differently or absent depending on the loading mechanism

Documentation Index