The Table Index Matcher node searches a Table Index Object using values provided in a second input table. Each row of the input table represents a search query that can span multiple indexed fields.
Users define the matching logic by dynamically adding, removing, reordering, and configuring field mappings between input columns and indexed fields. For each mapped field, users can specify:

• The indexed field to search
  Defines which indexed column inside the Multi-Field Index Object is evaluated against the selected input column.
• The relative importance (weight)
  Controls how strongly this field contributes to the overall multi-field match evaluation. Higher weights increase influence, lower weights reduce impact.
• Minimum acceptable quality
  Specifies the threshold (typically 0–100) that a match must reach for the field condition to be considered valid. A value of 100 represents an exact match.
• Filter checkbox
  Intended to control whether this field acts strictly as a filtering constraint or primarily contributes to scoring logic.
• Mode
  Defines the comparison strategy used for this field. Available modes depend on the data type.
  String Comparison Modes
  Exact – Requires full equality between values (quality = 100). No tolerance is applied.
  Approx – Uses Levenshtein-based similarity to allow tolerant matching in the presence of typos, spelling variations, or minor differences.
  Detect – Checks whether the input value appears within the indexed value (substring-style containment).
  Complete – Checks whether the indexed value, or a relevant part of it, exists within the input value.
  Numeric Comparison Modes
  Exact – Requires strict equality between numeric values.
  Approx – Compares numeric values based on their distance and adjusts the match quality according to how close they are. Higher similarity results in higher quality.
  Greater – Matches when the indexed value is greater than the input value.
  Lower – Matches when the indexed value is lower than the input value.

The node evaluates all configured field conditions deterministically and returns matching records along with match-related metadata. The output is designed for downstream analysis, entity resolution, and decision workflows.

Options

Select Settings Group

Allowing the user to navigate through different sections of the configuration options

• Search
• Output

Recall Columns

• Input Column: Selects the column from the input table that contains the search values. Each row in this column represents a query value used to search the multi-field index.
• Indexed Column: Selects the corresponding field inside the Table Index to search against. This determines which indexed attribute (e.g. email, name, city) the input values are matched to.
• Weight: Defines the importance of this field in the overall matching result. Higher weights increase the influence of this field when ranking matching records. Use this to prioritize more reliable fields (e.g. email over name).
• Minimum Quality: Threshold that a match must reach for the field condition to be considered valid. A value of 100 represents an exact match.
• Filter: Controls whether this field acts as a strict filtering constraint. If the values in this column do not match, the row is removed.
• Mode: Defines how values in the Input Column are compared against values in the Indexed Column during the matching process. The available comparison modes depend on the data type of the column.
  
  Comparison Modes for String Columns
  
  Approx
  Performs similarity-based matching using a Levenshtein distance algorithm. This mode enables tolerant matching by allowing small differences such as typos, spelling variations, or minor formatting inconsistencies. It is suitable when exact equality is not required.
  Detect
  Checks whether the value in the Input Column appears within the indexed value as a substring. This mode is useful when the input represents a partial query that should be contained within the indexed field.
  Example:
  Input = “John”
  Indexed = “John Michael Smith”
  → Match (Detect)
  Complete
  Checks whether the indexed value, or a meaningful portion of it, exists within the value of the Input Column. This mode is typically used when the indexed value represents a component of a larger input value.
  Example:
  Input = “John Michael Smith”
  Indexed = “Michael Smith”
  → Match (Complete)
  Exact
  Requires exact equality between the Input Column and the Indexed Column. Only matches (matching quality = 100) are accepted, and no fuzzy or approximate tolerance is applied.
  
  Comparison Modes for Numeric Columns
  
  Approximate
  Compares two numeric values by measuring the distance between them and adjusts the match quality based on how close they are. The smaller the difference between the values, the higher the match quality.
  Greater
  Returns matches where the indexed value is greater than the input value. This enables threshold-based filtering or range comparisons.
  Lower
  Returns matches where the indexed value is lower than the input value. Suitable for upper-bound filtering or range checks.
  Exact
  Requires strict equality between the input and indexed numeric values, with no tolerance for deviation.

Min. Match Value

Minimum total Match Value of indexed records

Max. Quality Spread

Maximum difference to best Match Value

Maximum Number of Search Results

Result Columns

Select columns from the index that are shown in the output.

Add Match Qualities of Search Columns

Append columns showing the match quality of each configured Search Column

Views

This node has no views

Workflows

• Address Matching with Multi-Field IndexerKNIME Hub
• Customer Identification via Email Content Matching (Data App)KNIME Hub
• Identity Resolution in Customer Data with Table IndexerKNIME Hub

Links

• No links available

Developers