Dataset Filter Configuration

This page describes how to use Dataset Filter in PI data ingestion tasks, including the applicable scenarios, wildcard syntax, and common examples for the three filter units: point, element, and template.

1. What Is Dataset Filter

When creating a PI or PI backfill task in Explorer, you can fill in the Dataset Filter in the Data Model Configuration area to filter PI Points or AF Elements before generating the model configuration file.

• To sync all points or all template elements, use the default configuration.
• To sync only points, elements, or templates whose names match specific patterns, fill in the Dataset Filter before clicking Download Default Configuration.

2. Relationship Among System Configuration, Data Model, and Dataset Filter

The available Dataset Filter options depend on the System Configuration (connection mode) and Data Model.

System Configuration	Data Model	Dataset Filter Unit	Description
PI Data Archive Only	Single-column model	point	Filter by PI Point name directly
PI Data Archive + AF Server	Single-column model	element / template	Filter by AF Element path or template name; each Point is still stored as a separate subtable
PI Data Archive + AF Server	Multi-column model	element / template	Filter by AF Element path or template name; each Element is stored as a separate subtable

Key points:

• PI Data Archive Only does not support AF, so only point filtering is available.
• Multi-column model relies on the AF asset framework, so it is only available in AF Server mode.
• Single-column model in AF Server mode uses element or template filtering; each Point is still stored as a separate subtable. This mode does not support point filtering because the single-column AF mode requires the Element → Attribute context to obtain metadata such as UOM and Element ownership.

3. Wildcard Syntax

All three Dataset Filter units support the following wildcards:

Wildcard	Meaning
*	Matches 0 or more characters
?	Matches exactly 1 character

4. Point Filter

Applicable when System Configuration = PI Data Archive Only, filtering by PI Point name.

4.1 Syntax

• Only name patterns are supported; path syntax is not supported.
• The expression is passed to the PI backend as-is; no * is appended automatically.

4.2 Examples

Input	Meaning
sinusoid	Exact match for the Point named sinusoid
CDT15*	Matches Points starting with CDT15
sinu?	Matches Points with sinu followed by any single character, e.g., sinus

5. Element Filter

Applicable when System Configuration = PI Data Archive + AF Server, filtering by AF Element name or path. It is the only filter type that supports path matching.

5.1 Path Matching Rules

• Paths are written relative to the AF database root; do not include the AF Database name.
• Leading \ is optional.
• The rule depends on whether the expression contains \:
  • Without \: the entire string is used as the element name pattern and matched across the whole database.
  • With \: split at the last \.
    • The left part is used as the Root: path (parent path).
    • The right part is used as the Name: pattern (empty means *).
    • Only leaf Elements (elements without child elements) are returned.

5.2 Examples

Input	Meaning
Meter_100001	Exact match for the Element named Meter_100001
Meter_10000*	Matches Elements starting with Meter_10000
Meter_100000?	Matches Elements with Meter_100000 followed by a single character
\California\San Diego\	Matches all leaf Elements under California\San Diego
\California\Meter_*	Matches leaf Elements under the California subtree whose names start with Meter_
\California\San Diego\Meter_10000?	Matches leaf Elements under the specified path whose names start with Meter_10000 and end with a single character

5.3 Notes

1. When selecting a path, the path must end with \. For example, \California\San Diego is parsed as looking for a leaf Element named San Diego under California; since San Diego is usually a grouping node, the result may be empty. The correct form is \California\San Diego\.
2. Each segment in the path must be an exact, real node name; path segments themselves do not support * or ? wildcards.
3. Do not include the AF Database name in the path. Paths are relative to the database root; including the database name will cause an error.
4. Path mode returns only leaf Elements. PI Points attached to non-leaf grouping nodes are not collected.

6. Template Filter

Applicable when System Configuration = PI Data Archive + AF Server, filtering by AF Element Template name.

6.1 Syntax

• Only template name patterns are supported; path syntax is not supported.
• The expression is passed to the PI backend as-is; no * is appended automatically.
• After a template is matched, taosX automatically collects all Elements that use that template.

6.2 Examples

Input	Meaning
MeterBasic	Exact match for the template named MeterBasic
Meter*	Matches templates starting with Meter
Meter?	Matches templates with Meter followed by a single character

7. Relationship Between Dataset Filter and CSV Configuration File

Dataset Filter only affects the range of points or elements generated when you click Download Default Configuration. After downloading, the CSV model configuration file contains:

• Multi-column model: the Filter row stores the Dataset Filter value.
• Single-column model: specific PI Points are listed in the point mapping section; you can manually delete unwanted Point rows for further trimming.

8. Common Misconceptions

Misconception	Clarification
Single-column model in AF Server mode supports point filtering	Not supported. The single-column AF mode requires the Element → Attribute context to obtain UOM, Element ownership, and other metadata; using point directly would lose this metadata.
Path segments in Element Filter can use * wildcards	Not supported. * / ? can only be used in the last segment (the name).
The AF Database name needs to be included in the path	No. Paths are relative to the database root; including the database name causes an error.
point / template support path syntax	Not supported. Only element supports path matching.