Skip to content

Predicted By

Definition

Asserts that the actual value of a field falls within an expected predicted range.

In-Depth Overview

The Predicted By rule is used to verify whether the actual values of a specific field align with a set of expected values that are derived from a prediction expression. This expression could be a mathematical formula, statistical calculation, or any other valid predictive logic.

Field Scope

Single: The rule evaluates a single specified field.

Accepted Fields

Type
Integral
Fractional
Date
Timestamp

General Properties

Name Supported
Filter
Allows the targeting of specific data based on conditions
Coverage Customization
Allows adjusting the percentage of records that must meet the rule's conditions

The filter allows you to define a subset of data upon which the rule will operate.

It requires a valid Spark SQL expression that determines the criteria rows in the DataFrame should meet. This means the expression specifies which rows the DataFrame should include based on those criteria. Since it's applied directly to the Spark DataFrame, traditional SQL constructs like WHERE clauses are not supported.

Examples

Direct Conditions

Simply specify the condition you want to be met.

Correct usage
O_TOTALPRICE > 1000
C_MKTSEGMENT = 'BUILDING'
Incorrect usage
WHERE O_TOTALPRICE > 1000
WHERE C_MKTSEGMENT = 'BUILDING'

Combining Conditions

Combine multiple conditions using logical operators like AND and OR.

Correct usage
O_ORDERPRIORITY = '1-URGENT' AND O_ORDERSTATUS = 'O'
(L_SHIPDATE = '1998-09-02' OR L_RECEIPTDATE = '1998-09-01') AND L_RETURNFLAG = 'R'
Incorrect usage
WHERE O_ORDERPRIORITY = '1-URGENT' AND O_ORDERSTATUS = 'O'
O_TOTALPRICE > 1000, O_ORDERSTATUS = 'O'

Utilizing Functions

Leverage Spark SQL functions to refine and enhance your conditions.

Correct usage
RIGHT(
    O_ORDERPRIORITY,
    LENGTH(O_ORDERPRIORITY) - INSTR('-', O_ORDERPRIORITY)
) = 'URGENT'
LEVENSHTEIN(C_NAME, 'Supplier#000000001') < 7
Incorrect usage
RIGHT(
    O_ORDERPRIORITY,
    LENGTH(O_ORDERPRIORITY) - CHARINDEX('-', O_ORDERPRIORITY)
) = 'URGENT'
EDITDISTANCE(C_NAME, 'Supplier#000000001') < 7

Using scan-time variables

To refer to the current dataframe being analyzed, use the reserved dynamic variable {{_qualytics_self}}.

Correct usage
O_ORDERSTATUS IN (
    SELECT DISTINCT O_ORDERSTATUS
    FROM {{_qualytics_self}}
    WHERE O_TOTALPRICE > 1000
)
Incorrect usage
O_ORDERSTATUS IN (
    SELECT DISTINCT O_ORDERSTATUS
    FROM ORDERS
    WHERE O_TOTALPRICE > 1000
)

While subqueries can be useful, their application within filters in our context has limitations. For example, directly referencing other containers or the broader target container in such subqueries is not supported. Attempting to do so will result in an error.

Important Note on {{_qualytics_self}}

The {{_qualytics_self}} keyword refers to the dataframe that's currently under examination. In the context of a full scan, this variable represents the entire target container. However, during incremental scans, it only reflects a subset of the target container, capturing just the incremental data. It's crucial to recognize that in such scenarios, using {{_qualytics_self}} may not encompass all entries from the target container.

Specific Properties

Determines if the actual value of a field falls within an expected predicted range.

Name Description
Expression
The prediction expression or formula for the field.
Tolerance
The allowed deviation from the predicted value.
Note

The tolerance level must be defined to allow a permissible range of deviation from the predicted values.

Here’s a simple breakdown:

  • An expression predicts what the value of a field should be.
  • A tolerance value specifies how much deviation from the predicted value is acceptable.
  • The actual value is then compared against the range defined by the predicted value ± tolerance.

Anomaly Types

Type Supported
Record
Flag inconsistencies at the row level
Shape
Flag inconsistencies in the overall patterns and distributions of a field

Example

Objective: Ensure that the discount (L_DISCOUNT) in the LINEITEM table is calculated correctly based on the actual price (L_EXTENDEDPRICE). A correct discount should be approximately 8% less than the actual price, within a tolerance of ±2.

Sample Data

L_ORDERKEY L_LINENUMBER L_EXTENDEDPRICE L_DISCOUNT
1 1 100 8
2 1 100 12
3 1 100 9
Inputs
  • Expression: L_EXTENDEDPRICE × 0.08
  • Tolerance: 2

Anomaly Explanation

For the entry with L_ORDERKEY 2, the discount is 12, which is outside of the computed range. Based on an 8% expected discount with a tolerance of ±2, the discount should be between 6 and 10 (calculated from the actual price of 100). Therefore, this record is marked as anomalous.

graph TD
A[Start] --> B[Retrieve L_EXTENDEDPRICE and L_DISCOUNT]
B --> C{Is Discount within Predicted Range?}
C -->|Yes| D[Move to Next Record/End]
C -->|No| E[Mark as Anomalous]
E --> D
-- An illustrative SQL query demonstrating the rule applied to example dataset(s).
select
    l_orderkey,
    l_linenumber,
    l_extendedprice,
    l_discount
from lineitem 
where
    l_discount not between l_extendedprice * 0.06 and l_extendedprice * 0.10;

Potential Violation Messages

Record Anomaly

The L_DISCOUNT value of '12' is not within the predicted range defined by L_EXTENDEDPRICE * 0.08 +/- 2.0

Shape Anomaly

In L_DISCOUNT, 33.333% of 3 filtered records (1) are not within the predicted range defined by L_EXTENDEDPRICE * 0.08 +/- 2.0


Last update: February 27, 2024