AO Match Building Block

Business name

(Required) Company name used for matching.

• Default: blank

Business name 2

(Optional) Alternate company name that can also be used for cross-field matching (for example, Lotus & IBM). Allows for different companies to be matched to either field.

• Default: blank

Business keyword

(Optional) Firm keyword to allow for matching to be qualified based on special field (for example, BCBS of MA vs. BCBS of ME) The ME/MA would be in its own field.

• Default: blank

Name type

Select input name type, either Full name or Parsed name.

• Default: full name

Name

If Contact type is Full name, the name field.

• Default: blank

First name

If Contact type is Parsed name, given name (John A Smith Jr).

• Default: blank

Middle name

If Contact type is Parsed name, middle name (John A Smith Jr).

• Default: blank

Last name

If Contact type is Parsed name, surname (John A Smith Jr).

• Default: blank

Suffix

If Contact type is Parsed name, generation name (John A Smith Jr).

• Default: blank

Gender

Gender. Must be Male, Female, or blank (unknown or indeterminate).

• Default: blank

Record ID

Optional. Field containing the unique record ID.

• Default: blank

Business score

Match threshold for business name field after any optional business adjustments (described below) are taking into account.

• Default: Medium-Tight (80)

Match nicknames

Allows for personal names in a firm to be standardized. For example, in Liz Smith Enterprises versus Elizabeth Smith Enterprises, Liz and Elizabeth would be treated as identical.

• Default: no

Initials

Recover match points if Initial matches a name (for example, J Robin Smith Inc versus Jonathon Robert Smith, Inc).

• Default: Treat as Similar (75)

Acronyms

Assign acronym matches a specific score for their part of the match (for example, International Machine Parts versus IMP).

• Default: Treat as Similar (75)

Abbreviations

Recover match points due to an abbreviation identified by pattern rather than known value (for example, Halbert Construction Contractors versus Hlbrt Construction Contractors).

• Default: Treat as Similar (75)

Missing words

Recover match points due to missing or disjoint words (for example, Halston Construction Contractors versus Halston Contractors).

• Default: Treat as Similar (75)

Word match threshold

Set minimum similarity threshold to consider any pair of words "the same."

• Default: (61)

Keyword score

Match threshold for business keyword field after optional business Keyword adjustments (described below) are taken into account.

• Default: blank

Match blank keyword

Specifies method for matching blank business keyword input field. Options are:

• Blanks Never Match: if either or both records have blank field, they will not match.

• Blank vs. Blank Only: (default) if both records have blank field, they will match. If only one is blank, they will not match.

• Blank vs. Non Blank Only: if only one record has blank field, they will match. If both have blank field, they will not match.

• Both One Blank and Both Blanks Match: matches either case; one or both records with a blank field for Business keyword.

Match abbreviations

Allows for variations in the business keyword field (for example, MISS vs. MISSISSIPPI to be considered an exact match as a case of abbreviation).

• Default: no

Ethnic nickname match

If selected, matches less common, but valid nicknames (such as Sean/John). Unwanted nicknames can be removed by adding a "remove" entry to the Name alias table.

• Default: no

Match gender

If selected, records with two different genders (no matter how close) will never match (for example, Alexander versus Alexandra). If a full name is used instead of parsed names or a gender field is not used, Data Management will attempt to internally generate one for matching purposes.

• Default: no

Ethnic nickname match

If selected, matches records on Last Name and Address. If selecting more than one match criteria, records must match on Resident to be compared as an Individual.

• Default: no

Match gender

Select this if you suspect that your records may have First name and Last name reversed.

• Default: no

Gender reversal

Defines how gender is handled in records where First Name and Last Name are reversed. Options are:

• Exclude: (default) existing gender is ignored; male records can match female records. For example, David Marie (M) would match Marie David (F).

• Include: existing gender is used for matching. For example. David Marie (M) would not match Marie David (F), even though the reversed text is identical.

• Regenderize: the reversed First Name (formerly the Last Name) is assigned a gender. "Male" records may still match "female records," depending on the last name: Mark David vs. Mary David.

Fix reversed first/last all recs

Select this if you selected Fix reversed first/last and you also want to fix records with an internal dedupe flag set to N.

• Default: no

Match first/middle

Select to enable cross comparison of first name against middle name.

• Default: no

Match first/initial

Select to enable cross comparison of first name against initial.

• Default: yes

Match middle/initial

Select to enable cross comparison of the middle name against initial.

• Default: yes

Ignore middle

Select to ignore middle name in name comparisons.

• Default: no

Fix reversed first/last all recs

Select to compare female records using only First Name (ignoring Last Name).

• Default: no

Match first/middle

Match threshold for First Name.

• Default: Medium (74)

Match first/initial

Match threshold for Middle Name.

• Default: Exact (100)

Match middle/initial

Match threshold for Last Name.

• Default: Medium (74)

"Never Match" override

If selected, use a second input to define "never match" pairs (pairs of record IDs that should never be matched). This input must contain two fields, ID1 and ID2.

• Default: no

This option operates at the record-comparison level, not the record-grouping level. So if you have three records with IDs {1,2,3} that all match each other, and inject "never match 1-3" using the never-match input, the records will still group due to the transitivity of matching 1-2 and 2-3.

ID1, ID2

If "Never Match" override is selected, the fields containing IDs for the "Never Match" pairs.

• Default: blank

Segment address data by

Specifies method for defining sort and comparison minimums for address data. Options are:

• FIELD 1—ALL

• FIELD 1—PARTIAL

• FIELD 1—AS EMAIL—DOMAIN

• FIELD 1—AS EMAIL—USER NAME

• FIELD 1—AS PHONE—LAST 7

• CUSTOM (SPECIAL FIELD)

The default is ZIP.

Custom segment

Optional. If Segment data by is Custom, the field containing the segment key.

• Default: blank

Partial segment chars

If you select FIELD 1—PARTIAL segmentation, define the number of characters to use from the field/column.

• Default: 1

Max segment size

This value controls the maximum number of records compared in a single segment, to prevent the compare process from running forever when segmentation is poorly defined. By default this allows for nearly-unlimited segment size. If you want to limit segment size to avoid runaway computation, potentially at the expense of missing a few record matches, set this to a lower value like 1000. Typically you can reduce this value unless you are matching within a very large segment like STATE.

• Default: 99999

Match segment with same value

The Match Building Block normally excludes from matching any group with the same value (111, 222, and so on). Enable this option when using with a value when it is permissible for a sequential value to exist.

• Default: no

Data sorted by segment

Enable this option if your data is already sorted by the segment field(s). You'll improve execution speed by avoiding re-sorting the data.

The data is sorted lexically rather than numerically, so numeric data must have leading zeros.

• Default: no

Custom segment

Optional. If you select CUSTOM segmentation, specify the field containing the segment key. This is useful if the same type of data (i.e. Home Phone, Work Phone, Cell Phone) is in multiple fields and you want to cross-compare.

• Default: blank

Max segment size

This value controls the maximum number of records compared in a single segment, to prevent the compare process from running forever when segmentation is poorly defined. By default this allows for nearly-unlimited segment size. If you want to limit segment size to avoid runaway computation, potentially at the expense of missing a few record matches, set this to a lower value like 1000. Typically you can reduce this value unless you are matching within a very large segment like STATE.

• Default: 99999

Data sorted by segment

Enable this option if your data is already sorted by the segment field(s). You'll improve execution speed by avoiding re-sorting the data.

The data is sorted lexically rather than numerically, so numeric data must have leading zeros.

• Default: no

Output match score

If selected, outputs the overall score from the match records as a percentage between 1—100.

• Default: no

Match score

Field for match score.

• Default: MATCH_SCORE

Match ID

The Match ID (or Group ID) generated by the matching process. This defines the match groups.

• Default: MATCH_GROUP

Source

Field containing the logical description for input data source. This is usually defined in AO Define Source.

• Default: blank

Internal dedupe flag

Field containing a Y/N flag indicating whether or not data from a particular source should be compared against itself (deduped) or solely against other sources. As a general rule, master databases are not internally deduped whereas update files are.

• Default: blank

Compare sources not internally deduped

If a matching process has more than one source with the internal dedupe field set to "N", selecting this will compare the two sources.

• Default: yes

Parallelism level

Set to the lesser of the number of CPU cores on the Execution Server, or the number of threads configured in the project in which the macro is embedded.

• Default: 1

Optimize for large segments

If you receive warnings like "Window Compare segment size for value (06828EAS3135) has exceeded 2000," enable this option. Comparing large candidate groups may reduce matching efficiency. Selecting this option increases the number of records that can be sent to a matching process without slowing processing.

• Default: no

Name alias table

Path and name of optional DLD table used to augment or override the alias values defined within the macro. The table must be of the following form.

Where ALIAS and STANDARD are Text fields and REMOVE is Boolean. In the example above, Peg and Margie are defined as new aliases for Margaret (a blank REMOVE field is treated as FALSE), while the TRUE value in the REMOVE field explicitly suppresses Jon as an alias for John.

• Default: blank

Business alias table

Path and name of optional DLD table used to augment or override the alias values defined within the macro. The table must be of the form below.

Where ALIAS and STANDARD are Text fields and REMOVE is Boolean. In the example above, Paving and Hotmix are defined as new aliases for Asphalt (a blank REMOVE field is treated as FALSE), while the TRUE value in the REMOVE field explicitly suppresses Reproduction as an alias for Copying.

• Default: blank

Business noise table

Path and name of optional DLD table used to add additional "noise" words for Firm matching. The table must be a single-column DLD table of the form below.

Where WORDS is a Text field. The example above shows data that shouldn’t contribute to the match score because it is contextually meaningless. If you are matching financial institutions, word BANK could be a "noise" word.

• Default: blank

Field 1

(Required) Map a field for segmentation and/or to match for this macro.

• Default: none

Score

Match threshold for Field 1 field.

• Default: Tight (88)

Blank matching

Specifies how blank Field 1 input field is matched. Options are:

• NONE: if either or both records have blank field, they will not match.

• BOTH: (default) if both records have blank field, they will match. If only one is blank, they will not match.

• ONE: only one record has blank field, they will match. If both have blank field, they will not match.

• ALL: matches either case—one or both records with a blank field for Field 1.

Comparison kind

Specifies the field comparison method. Options are:

• Positional: compares each character position within the two records.

• Edit Distance: (default) compares the fields of the two records using an algorithm that counts how many "mistakes" were made to transform one field value into the other.

• Word-by-Word: compares the fields of two records by splitting the text of each field into words (punctuation and spaces are dropped) and then comparing the words one at a time. This method is often used for business-name comparison where word order is not as important.

Sensitivity

Differentiates values that are close together. For example:

With Sensitivity set to 1:

With Sensitivity set to 10:

• Default: 1

Zeros as blanks

Specifies that values of zero are treated as blanks for the purposes of Blank matching.

• Default: on

Numeric threshold

If selected, lets you specify a secondary match threshold (Numeric Minimum Score) for the digits contained in the match values. A second match is performed on the digits of both values using an edit distance algorithm. The numeric score is computed by extracting all the digits, and applying the following rules:

• If both digit sets are blank, it is a match.

• If only one digit set is blank, it is a non-match.

• If both digit sets are non-blank, they are compared using Edit Distance Qwerty rules.

If this match fails to meet the threshold, the entire match fails. Use this if your field contains both digits and non-digits, but the digits are more critical to the match.

• Default: no

Minimum Score

Numeric secondary match threshold, as described above.

• Default: Exact (100)

Initials

Recover match points if Initial matches a name (for example, J Robin Smith Inc versus Jonathon Robert Smith, Inc).

• Default: Treat as Different (0)

Abbreviations

Recover match points due to an abbreviation identified by pattern rather than known value (for example, Halbert Construction Contractors versus Hlbrt Construction Contractors).

• Default: Treat as Different (0)

Missing words

Recover match points due to missing or disjoint words (for example, Halston Construction Contractors versus Halston Contractors).

• Default: Treat as Different (0)

Acronyms

Assign acronym matches a specific score for their part of the match (for example, International Machine Parts versus IMP).

• Default: Treat as Different (0)

Word match threshold

Set minimum similarity threshold to consider any pair of words "the same."

• Default: Treat as Different (0)