Text Metric

The TEXT-METRIC directive provides a metric (from 0 to 1) measuring the difference between two sequence of characters, using a specified method of measuring the distance between strings.

Syntax

text-metric <method> <column-1> <column-2> <destination>
  • <method> specifies the method to be used to measure the distance between the strings of <column-1> and <column-2>
  • <destination> is the column the resulting difference will be stored in. If it exists, it will be overwritten. If it does not exist, it will be created.

Note: If either or both of the two columns do not exist, no error will be returned, and the destination column will still be created or overwritten.

Usage Notes

The value of the metric is always between 0 and 1.

These distance measure methods are supported:

  • block-distance
  • block
  • cosine
  • damerau-levenshtein
  • dice
  • euclidean
  • generalized-jaccard
  • identity
  • jaccard
  • jaro
  • levenshtein
  • longest-common-subsequence
  • longest-common-substring
  • overlap-cofficient
  • simon-white

Example

Using this record as an example:

{
  "tweet1": "CheeseBurgers are God tellin us everything's gonna be cool...",
  "tweet2": "Beer is God's gift to humankind."
}

Applying this directive:

text-metric longest-common-subsequence tweet1 tweet2 distance

would result in this record:

{
  "tweet1": "CheeseBurgers are God tellin us everything's gonna be cool...",
  "tweet2": "Beer is God's gift to humankind."
  "destination": 0.26229507
}