ASDD standard profile set
As of the 11.0.0.0 release, administrators control when upgrades of the ASDD Standard Profile Set occur. For every Delphix Engine release, the ASDD Standard Profile Set will be made available to users in the Automated Sensitive Data Discovery directory folder on the Delphix Download site.
A README is also present, which includes a change.log and usage instructions. Additionally, as of 11.0.0.0, the contents of the ASDD Standard Profile Set may be customized. To recover or upgrade the ASDD Standard profile set, see the ASDD Profile Set Import and Export article for more information on the import API endpoint.
For reference, you can download or expand the JSON listing below, which contains the full configuration for all classifiers present in the ASDD Standard Profile Set as of the 21.0.0.0 release. The profile set for individual releases can be found in the downloads section.
ASDD Standard Profile Set JSON classifiers listing
[
{
"domain": "ACCOUNT_NO",
"name": "Account Number - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(account|accnt|acct)_?-? ?(number|num|nbr|no|user))($|[ _-])" }
],
"rejectStrength": 0.0
}
},
{
"domain": "ACCOUNT_NO",
"name": "Account Number - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "5",
"typeName": "String" },
{
"minimumLength": "5",
"typeName": "Number" }
],
"matchAutoIncrementingColumn": true
}
},
{
"domain": "ADDRESS",
"name": "Address Line 1 - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>((st(reet)?_?-? ?addr(ess)?)|street?|(?<!email[-_ ]?|ip[-_ ]?|web[-_ ]?|url[-_ ]?)addr(ess)?)_?-? ?((l(i)?n(e)?)? ?_?(1|))?)" }
],
"rejectStrength": 0.0
}
},
{
"domain": "ADDRESS",
"name": "Address Line 1 - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "20",
"typeName": "String" }
]
}
},
{
"domain": "ADDRESS",
"name": "Address Line 1 - Regex",
"type": "REGEX",
"properties": {
"dataPatterns": [
{
"regex": "(?i)(.*[\\s]+b(ou)?l(e)?v(ar)?d[\\s]*.*)|(.*[\\s]+st(reet)?[\\s]*.*)|(.*[\\s]+ave[.]?(nue)?[\\s]*.*)|(.*[\\s]+r(oa)?d[\\s]*.*)|(.*[\\s]+l(a)?n(e)?[\\s]*.*)|(.*[\\s]+cir(cle)?[\\s]*.*)|(.*[\\s]+dr[.]?(ive)?[\\s]*.*)|(.*[\\s]+h(igh)?w(a)?y[\\s]*.*)|(.*[\\s]+r(ou)?t(e)?[\\s]*.*)|(.*[\\s]+c(our)?t[\\s]*.*)|(.*[\\s]+way[\\s]*.*)",
"matchStrength": 0.7,
"allowPartialMatch": false
}
],
"rejectStrength": 0.1
}
},
{
"domain": "ADDRESS_LINE2",
"name": "Address Line 2 - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>((st(reet)?_?-? ?addr(ess)?)|street?|(?<!email[-_ ]?|ip[-_ ]?|web[-_ ]?|url[-_ ]?)addr(ess)?)_?-? ?((l(i)?n(e)?)? ?_?([2-9]))?)" }
],
"rejectStrength": 0.0
}
},
{
"domain": "ADDRESS_LINE2",
"name": "Address Line 2 - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "20",
"typeName": "String" }
]
}
},
{
"domain": "ADDRESS_LINE2",
"name": "Address Line 2 - Regex",
"type": "REGEX",
"properties": {
"dataPatterns": [
{
"regex": "(?i)(.*[\\s]*ap(ar)?t(ment)?[.\\s]+.*)|(.*[\\s]*s(ui)?te[.\\s]+.*)|(c(are)?[\\s]*[\\\\]?[/]?o(f)?[.\\s]+.*)|(.*[\\s]*b(ui)?ld(in)?g[.\\s]+.*)|(.*[\\s]*fl(oor)?[.\\s]+.*)|(.*[\\s]*r(oo)?m[.\\s]+.*)|(.*[\\s]*unit[.\\s]+.*)",
"matchStrength": 0.7,
"allowPartialMatch": false
}
],
"rejectStrength": 0.1
}
},
{
"domain": "AGE",
"name": "Age - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(^|[ _-])(age)[ _-]?(group|grp|number|num|nbr|no)?)($|[ _-])" }
],
"rejectStrength": 1.0
}
},
{
"domain": "AGE",
"name": "Age - Regex",
"type": "REGEX",
"properties": {
"dataPatterns": [
{
"note": "Number range between 01 --> 99",
"regex": "\\d{1,2}",
"matchStrength": 0.1,
"allowPartialMatch": false
}
],
"rejectStrength": 0.1
}
},
{
"domain": "AGE",
"name": "Age - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "2",
"typeName": "Number" },
{
"minimumLength": "2",
"typeName": "String" }
]
}
},
{
"domain": "BANK_ACCOUNT_NO",
"name": "Bank Account Number - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(bank_?-? ?)?(account|accnt|acct)_?-? ?(number|num|nbr|no))($|[ _-])" }
],
"rejectStrength": 0.0
}
},
{
"domain": "BANK_ACCOUNT_NO",
"name": "Bank Account Number - Regex",
"type": "REGEX",
"properties": {
"dataPatterns": [
{
"regex": "\\d{5,17}",
"matchStrength": 0.05,
"allowPartialMatch": false
}
],
"rejectStrength": 0.1
}
},
{
"domain": "BANK_ACCOUNT_NO",
"name": "Bank Account Number - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "5",
"typeName": "Number" }
],
"matchAutoIncrementingColumn": true
}
},
{
"domain": "BENEFICIARY_NO",
"name": "Beneficiary Number - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(bene(ficiary)?_?-? ?(number|num|nbr|no|id)))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "BENEFICIARY_NO",
"name": "Beneficiary Number - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "5",
"typeName": "Number" },
{
"minimumLength": "10",
"typeName": "String" }
]
}
},
{
"domain": "BIOMETRIC",
"name": "Biometric - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(biometric))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "BIOMETRIC",
"name": "Biometric - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "10",
"typeName": "String" },
{
"minimumLength": "0",
"typeName": "Binary" }
]
}
},
{
"domain": "BLOOD_TYPE",
"name": "Blood Type - List",
"type": "LIST",
"properties": {
"valueLists": [
{
"file": "file://blood_types.txt",
"matchStrength": 1.0
}
]
}
},
{
"domain": "BLOOD_TYPE",
"name": "Blood Type - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?>blood[-_ ]?(group|grp|type))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "BLOOD_TYPE",
"name": "Blood Type - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "2",
"typeName": "String" }
]
}
},
{
"domain": "CERTIFICATE_NO",
"name": "Certificate Number - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(^|[-_ ])cert(ificate)?_?-? ?(number|num|nbr|no|id))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "CERTIFICATE_NO",
"name": "Certificate Number - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "10",
"typeName": "String" },
{
"minimumLength": "5",
"typeName": "Number" }
]
}
},
{
"domain": "CITY",
"name": "City - List",
"type": "LIST",
"properties": {
"valueLists": [
{
"file": "file://us_cities.txt",
"matchStrength": 1.0
}
]
}
},
{
"domain": "CITY",
"name": "City - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>^(home_?-? ?city|city))" },
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>^(address_?-? ?city|city|city_?-? ?address))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "CITY",
"name": "City - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "10",
"typeName": "String" }
]
}
},
{
"domain": "COMPANY_NAME",
"name": "Company Name - List",
"type": "LIST",
"properties": {
"valueLists": [
{
"file": "file://companies.txt",
"matchStrength": 1.0
}
]
}
},
{
"domain": "COMPANY_NAME",
"name": "Company Name - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?>(comp|company|org|organization|employer)[-_ ]?(name|nm))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "COMPANY_NAME",
"name": "Company Name - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "6",
"typeName": "String" }
]
}
},
{
"domain": "COUNTRY",
"name": "Country - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)country",
"allowPartialMatch": false
}
],
"rejectStrength": 0.0
}
},
{
"domain": "COUNTRY",
"name": "Country - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "15",
"typeName": "String" }
]
}
},
{
"domain": "COUNTRY",
"name": "Country - List",
"type": "LIST",
"properties": {
"valueLists": [
{
"file": "file://countries.txt",
"matchStrength": 1.0
}
],
"rejectStrength": 0.5
}
},
{
"domain": "COUNTY",
"name": "County - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(county))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "COUNTY",
"name": "County - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "10",
"typeName": "String" }
]
}
},
{
"domain": "CREDIT CARD",
"name": "Credit Card Number - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(^cc|credit[ -_]?card)[ -_]?(number|num|nbr|no)?)" },
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>card[ -_]?(number|num|nbr|no))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "CREDIT CARD",
"name": "Credit Card Number - Regex",
"type": "REGEX",
"properties": {
"dataPatterns": [
{
"regex": "(?:3[47][0-9]{2}[0-9]{6}[0-9]{4})",
"allowPartialMatch": false,
"checksumType": "LUHN",
"dataCleanRegex": "[ -.]",
"matchStrength": 0.9
},
{
"regex": "(?:4[0-9]{12}(?:[0-9]{3})?(?:[0-9]{3})?)",
"allowPartialMatch": false,
"checksumType": "LUHN",
"dataCleanRegex": "[ -.]",
"matchStrength": 0.9
},
{
"regex": "(?:5[1-5][0-9]{2}|222[1-9]|22[3-9][0-9]|2[3-6][0-9]{2}|27[01][0-9]|2720)[0-9]{12}",
"allowPartialMatch": false,
"checksumType": "LUHN",
"dataCleanRegex": "[ -.]",
"matchStrength": 0.9
},
{
"regex": "(?:2131|1800|35[0-9]{3})[0-9]{11}",
"allowPartialMatch": false,
"checksumType": "LUHN",
"dataCleanRegex": "[ -.]",
"matchStrength": 0.9
},
{
"regex": "3(?:0[0-5,9]|6[0-9])[0-9]{11}|3[89][0-9]{12}?(?:[0-9]{1,3})?",
"allowPartialMatch": false,
"checksumType": "LUHN",
"dataCleanRegex": "[ -.]",
"matchStrength": 0.9
},
{
"regex": "6(?:(011|5[0-9][0-9])[0-9]{2}|4[4-9][0-9]{3}|2212[6-9]|221[3-9][0-9]|22[2-8][0-9]{2}|229[0-1][0-9]|2292[0-5])[0-9]{10}?(?:[0-9]{3})?",
"allowPartialMatch": false,
"checksumType": "LUHN",
"dataCleanRegex": "[ -.]",
"matchStrength": 0.9
}
],
"rejectStrength": 0.1
}
},
{
"domain": "CREDIT CARD",
"name": "Credit Card Number - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "15",
"typeName": "Number" },
{
"minimumLength": "15",
"typeName": "String" }
]
}
},
{
"domain": "CUSTOMER_NO",
"name": "Customer Number - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(cust(omer|mr)?) ?_?-?(num(ber)?|nbr|no))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "CUSTOMER_NO",
"name": "Customer Number - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "5",
"typeName": "String" },
{
"minimumLength": "5",
"typeName": "Number" }
]
}
},
{
"domain": "DOB",
"name": "Date of Birth - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>dob|dtofb|(day|date?|dt)_?-?(of)?_?(birth))" },
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>b(irth)?_?-? ?(date|day|dt))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "DOB",
"name": "Date of Birth - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "6",
"typeName": "String" },
{
"typeName": "Date" }
]
}
},
{
"domain": "DEPARTMENT",
"name": "Department - List",
"type": "LIST",
"properties": {
"valueLists": [
{
"file": "file://departments.txt",
"matchStrength": 1.0
}
],
"tokenizeInput": true,
"tokenizationDelimiter": " " }
},
{
"domain": "DEPARTMENT",
"name": "Department - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?>((dept|depa?rtme?nt)[-_ ]?(name|nm)))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "DEPARTMENT",
"name": "Department - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "5",
"typeName": "String" }
]
}
},
{
"domain": "DRIVING_LC",
"name": "Drivers License - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(drivers?|lic(ense)?)_?-? ?(number|num|nbr|no))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "DRIVING_LC",
"name": "Drivers License - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "10",
"typeName": "Number" },
{
"minimumLength": "10",
"typeName": "String" }
]
}
},
{
"domain": "DRIVING_LC",
"name": "Drivers License - Regex",
"type": "REGEX",
"properties": {
"dataPatterns": [
{
"note": "One alpha followed by digits (short): AZ 1+8, CA 1+7, HI 1+8, IN 1+9, KS 1+8, MA 1+9, MO 1+5..9, MT 1+9, NE 1+6..8, NV 1+8, OH 1+4..8, OK 1+9, OR 1+6..7, RI 1+6, UT 4-10, VA 1+8..11, VT 1+8, WV 1+6",
"regex": "[A-Z][0-9]{4,9}",
"matchStrength": 0.15,
"allowPartialMatch": false
},
{
"note": "One alpha followed by digits (long): FL 1+12, MD 1+12, MI 1+12, MN 1+12, NJ 1+14, UT 4-10, VA 1+8..11",
"regex": "[A-Z][0-9]{10,14}",
"matchStrength": 0.3,
"allowPartialMatch": false
},
{
"note": "One alpha followed by digits with dashes (short): KS 1+8, KY 1+8, VA 1+8..11",
"dataCleanRegex": "[-]",
"regex": "[A-Z][0-9]{8,9}",
"matchStrength": 0.15,
"allowPartialMatch": false
},
{
"note": "One alpha followed by digits with dashes (long): FL 1+12, IL 1+11 or 1+12, MD 1+12, MN 1+12, VA 1+8..11, WI 1+13",
"dataCleanRegex": "[-]",
"regex": "[A-Z][0-9]{10,13}",
"matchStrength": 0.3,
"allowPartialMatch": false
},
{
"note": "Two alpha followed by digits: OH 2+3..7, WV 2+7",
"regex": "[A-Z]{2}[0-9]{3,7}",
"matchStrength": 0.2,
"allowPartialMatch": false
},
{
"note": "Digits followed by alpha: ME 7+1, MO 8+2 or 9+1, VT 7+1",
"regex": "[0-9]{7}[A-Z]|[0-9]{8}[A-Z]{2}|[0-9]{9}[A-Z]",
"matchStrength": 0.2,
"allowPartialMatch": false
},
{
"note": "Digits and spaces: NM 3-3-3, NY 3-3-3, PA 2-3-3",
"regex": "[0-9]{2,3} [0-9]{3} [0-9]{3}",
"matchStrength": 0.1,
"allowPartialMatch": false
},
{
"note": "Digit-alpha-digit: IA 3+2+4, NH 2+3+5",
"regex": "[0-9]{3}[A-Z]{2}[0-9]{4}|[0-9]{2}[A-Z]{3}[0-9]{5}",
"matchStrength": 0.3,
"allowPartialMatch": false
},
{
"note": "Digits only: AL 7-8, AK 7, AZ 9, AR 9, CT 9, DE 7, DC 7, GA 9, HI 9, IA 9, LA 8, ME 7-8, MA 9, M0 9, MT 9 or 13-14, NV 9-10 or 12, NY 9, NM 8-9, NC 1-12, ND 9, OH 8, OK 9, RI 7, SC 5-11, SD 6-10 or 12, TN 7-9, TX 7-8, VA 9, WV 7, WY 9",
"regex": "([0-9]{6,14})",
"matchStrength": 0.0,
"allowPartialMatch": false
},
{
"note": "CO",
"regex": "[0-9]{2}-[0-9]{3}-[0-9]{4}",
"matchStrength": 0.1,
"allowPartialMatch": false
},
{
"note": "ID",
"regex": "[A-Z]{2}[0-9]{6}[A-Z]",
"matchStrength": 0.3,
"allowPartialMatch": false
},
{
"note": "IN",
"regex": "[0-9]{2}-[0-9]{4}-[0-9]{4}",
"matchStrength": 0.1,
"allowPartialMatch": false
},
{
"note": "MA - SA + 7 digits",
"regex": "SA[0-9]{7}",
"matchStrength": 0.5,
"allowPartialMatch": false
},
{
"note": "MI - 1 alpha + 12 digits with spaces",
"regex": "[A-Z]( [0-9]{3}){4}",
"matchStrength": 0.5,
"allowPartialMatch": false
},
{
"note": "MS - looks just like SSN with dashes",
"regex": "([0-9]{3}-[0-9]{2}-[0-9]{4})",
"matchStrength": 0.0,
"allowPartialMatch": false
},
{
"note": "MO - not covered elsewhere",
"regex": "[0-9]{3}[A-Z][0-9]{6}|[A-Z][0-9]{6}R",
"matchStrength": 0.3,
"allowPartialMatch": false
},
{
"note": "NH - NHL + 8 digits",
"regex": "NHL[0-9]{8}",
"matchStrength": 0.5,
"allowPartialMatch": false
},
{
"note": "ND - 3 letters, 6 digits with spaces",
"regex": "[A-Z]{3} ?[0-9]{2} ?[0-9]{4}",
"matchStrength": 0.3,
"allowPartialMatch": false
},
{
"note": "NJ 1 alpha + 14 digits with spaces",
"dataCleanRegex": "[-]",
"regex": "[A-Z][0-9]{4} [0-9]{5} [0-9]{5}",
"matchStrength": 0.4,
"allowPartialMatch": false
},
{
"note": "WA - old - very broad will match any 12 position alpha string",
"regex": "[A-Z]([A-Z]{4}|[A-Z]{3}[*]|[A-Z]{2}[*]{2}|[A-Z]{1}[*]{3})[A-Z]{2}[0-9A-Z]{5}",
"matchStrength": 0.0,
"allowPartialMatch": false
},
{
"note": "WA - new",
"regex": "WDL[0-9A-Z]{9}",
"matchStrength": 0.3,
"allowPartialMatch": false
},
{
"note": "WY - 9 digits with a dash",
"regex": "[0-9]{6}-[0-9]{3}",
"matchStrength": 0.1,
"allowPartialMatch": false
}
],
"rejectStrength": 0.75
}
},
{
"domain": "EMAIL",
"name": "Email Address - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(cust|customer|partner|home|private|def|default)_?-? ?(email)_?-? ?(address|)" },
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(email_?-? ?)(addr?e?s?s?)?)" }
],
"rejectStrength": 0.0
}
},
{
"domain": "EMAIL",
"name": "Email Address - Regex",
"type": "REGEX",
"properties": {
"dataPatterns": [
{
"regex": "[A-Z0-9.!#$%&'*+/=?^_{|}~-]{1,64}@(?=.{1,255}$)[A-Z0-9-]+(?:\\.[A-Z0-9-]+)*",
"matchStrength": 0.9,
"allowPartialMatch": false
}
],
"rejectStrength": 0.1
}
},
{
"domain": "EMAIL",
"name": "Email Address - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "20",
"typeName": "String" }
]
}
},
{
"domain": "ETHNICITY",
"name": "Ethnicity - List",
"type": "LIST",
"properties": {
"valueLists": [
{
"file": "file://ethnicities.txt",
"matchStrength": 1.0
}
]
}
},
{
"domain": "ETHNICITY",
"name": "Ethnicity - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?>ethnicity|ethnicities|^race)" }
],
"rejectStrength": 0.0
}
},
{
"domain": "ETHNICITY",
"name": "Ethnicity - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "7",
"typeName": "String" }
]
}
},
{
"domain": "FIRST_NAME",
"name": "First Name - List",
"type": "LIST",
"properties": {
"valueLists": [
{
"file": "file://us_first.txt",
"matchStrength": 1.0
},
{
"file": "file://de_first.txt",
"matchStrength": 1.0
},
{
"file": "file://ch_first.txt",
"matchStrength": 1.0
}
]
}
},
{
"domain": "FIRST_NAME",
"name": "First Name - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(mid(dle)?_?-? ?(na?me?))(_?-?user)?)" },
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(f(irst)?_?-? ?(na?me?))(_?-?user)?)" }
],
"rejectStrength": 0.0
}
},
{
"domain": "FIRST_NAME",
"name": "First Name - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "10",
"typeName": "String" }
]
}
},
{
"domain": "TEXT",
"name": "Text - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?>(free[-_ ]?text|comment|description|note|remark|text|user[-_ ]?input|user[-_ ]?text))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "TEXT",
"name": "Text - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "20",
"typeName": "String" }
]
}
},
{
"domain": "FULL_NAME",
"name": "Full Name - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>((fu?l?l|whole|user)([-_ ]*)?(na?me?)))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "FULL_NAME",
"name": "Full Name - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "20",
"typeName": "String" }
]
}
},
{
"domain": "FULL_NAME",
"name": "Full Name - List",
"type": "LIST",
"properties": {
"valueLists": [
{
"file": "file://us_first.txt",
"matchStrength": 0.7
},
{
"file": "file://de_first.txt",
"matchStrength": 0.7
},
{
"file": "file://ch_first.txt",
"matchStrength": 0.7
},
{
"file": "file://us_last.txt",
"matchStrength": 0.7
},
{
"file": "file://de_last.txt",
"matchStrength": 0.7
},
{
"file": "file://ch_last.txt",
"matchStrength": 0.7
}
],
"tokenizeInput": true,
"tokenizationDelimiter": " " }
},
{
"domain": "GENDER",
"name": "Gender - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(ge?nde?r|sex)[-_ ]?(type|identity)?)($|[-_ ])" }
],
"rejectStrength": 0.0
}
},
{
"domain": "GENDER",
"name": "Gender - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "5",
"typeName": "String" }
]
}
},
{
"domain": "GENDER",
"name": "Gender - List",
"type": "LIST",
"properties": {
"valueLists": [
{
"file": "file://genders.txt",
"matchStrength": 0.9
}
],
"rejectStrength": 0.1
}
},
{
"domain": "HIPAA_DATE",
"name": "HIPAA Date - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(adm(it|ission)?|tr(ea)?t(ment)?_?-?|ds|disc(h|harge))-?_? ?(date|day|dt))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "HIPAA_DATE",
"name": "HIPAA Date - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "6",
"typeName": "String" },
{
"typeName": "Date" }
]
}
},
{
"domain": "HOUSE_NO",
"name": "House Number - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?>(house?|plot)[_ -]?(number|num|nbr|no|id))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "HOUSE_NO",
"name": "House Number - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "1",
"typeName": "String" },
{
"minimumLength": "1",
"typeName": "Number" }
]
}
},
{
"domain": "IBAN",
"name": "IBAN - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?>(^|[-_ ])iban|(international|int|intl)[-_ ]?bank[-_ ]?acc(ount)?[-_ ]?(number|num|nbr|no))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "IBAN",
"name": "IBAN - Regex",
"type": "REGEX",
"properties": {
"dataPatterns": [
{
"note": "Fetched from: 'https://www.swift.com/standards/data-standards/iban-international-bank-account-number', List of countries in order for regex formats listed below: Andorra, United Arab Emirates (The), Albania, Austria, Azerbaijan, Bosnia and Herzegovina, Belgium, Bulgaria, Bahrain, Burundi, Brazil, Republic of Belarus, Switzerland, Costa Rica, Cyprus, Czechia, Germany, Djibouti, Denmark, Dominican Republic, Estonia, Egypt, Spain, Finland, Faroe Islands, France, United Kingdom, Georgia, Gibraltar, Greenland, Greece, Guatemala, Croatia, Hungary, Ireland, Israel, Iraq, Iceland, Italy, Jordan, Kuwait, Kazakhstan, Lebanon, Saint Lucia, Liechtenstein, Lithuania, Luxembourg, Latvia, Libya, Monaco, Moldova, Montenegro, Macedonia, Mauritania, Malta, Mauritius, Netherlands, Norway, Pakistan, Poland, Palestine, State of, Portugal, Qatar, Romania, Serbia, Russia, Saudi Arabia, Seychelles, Sudan, Sweden, Slovenia, Slovakia, San Marino, Somalia, Sao Tome and Principe, El Salvador, Timor-Leste, Tunisia, Turkey, Ukraine, Vatican City State, Virgin Islands, Kosovo",
"regex": "(AD[0-9]{10}[A-Z0-9]{12})|(AE[0-9]{21})|(AL[0-9]{10}[A-Z0-9]{16})|(AT[0-9]{18})|(AZ[0-9]{2}[A-Z]{4}[A-Z0-9]{20})|(BA[0-9]{18})|(BE[0-9]{14})|(BG[0-9]{2}[A-Z]{4}[0-9]{6}[A-Z0-9]{8})|(BH[0-9]{2}[A-Z]{4}[A-Z0-9]{14})|(BI[0-9]{25})|(BR[0-9]{25}[A-Z]{1}[A-Z0-9]{1})|(BY[0-9]{2}[A-Z0-9]{4}[0-9]{4}[A-Z0-9]{16})|(CH[0-9]{7}[A-Z0-9]{12})|(CR[0-9]{20})|(CY[0-9]{10}[A-Z0-9]{16})|(CZ[0-9]{22})|(DE[0-9]{20})|(DJ[0-9]{25})|(DK[0-9]{16})|(DO[0-9]{2}[A-Z0-9]{4}[0-9]{20})|(EE[0-9]{18})|(EG[0-9]{27})|(ES[0-9]{22})|(FI[0-9]{16})|(FO[0-9]{16})|(FR[0-9]{12}[A-Z0-9]{11}[0-9]{2})|(GB[0-9]{2}[A-Z]{4}[0-9]{14})|(GE[0-9]{2}[A-Z]{2}[0-9]{16})|(GI[0-9]{2}[A-Z]{4}[A-Z0-9]{15})|(GL[0-9]{16})|(GR[0-9]{9}[A-Z0-9]{16})|(GT[0-9]{2}[A-Z0-9]{24})|(HR[0-9]{19})|(HU[0-9]{26})|(IE[0-9]{2}[A-Z]{4}[0-9]{14})|(IL[0-9]{21})|(IQ[0-9]{2}[A-Z]{4}[0-9]{15})|(IS[0-9]{24})|(IT[0-9]{2}[A-Z]{1}[0-9]{10}[A-Z0-9]{12})|(JO[0-9]{2}[A-Z]{4}[0-9]{4}[A-Z0-9]{18})|(KW[0-9]{2}[A-Z]{4}[A-Z0-9]{22})|(KZ[0-9]{5}[A-Z0-9]{13})|(LB[0-9]{6}[A-Z0-9]{20})|(LC[0-9]{2}[A-Z]{4}[A-Z0-9]{24})|(LI[0-9]{7}[A-Z0-9]{12})|(LT[0-9]{18})|(LU[0-9]{5}[A-Z0-9]{13})|(LV[0-9]{2}[A-Z]{4}[A-Z0-9]{13})|(LY[0-9]{23})|(MC[0-9]{12}[A-Z0-9]{11}[0-9]{2})|(MD[0-9]{2}[A-Z0-9]{20})|(ME[0-9]{20})|(MK[0-9]{5}[A-Z0-9]{10}[0-9]{2})|(MR[0-9]{25})|(MT[0-9]{2}[A-Z]{4}[0-9]{5}[A-Z0-9]{18})|(MU[0-9]{2}[A-Z]{4}[0-9]{19}[A-Z]{3})|(NL[0-9]{2}[A-Z]{4}[0-9]{10})|(NO[0-9]{13})|(PK[0-9]{2}[A-Z]{4}[A-Z0-9]{16})|(PL[0-9]{26})|(PS[0-9]{2}[A-Z]{4}[A-Z0-9]{21})|(PT[0-9]{23})|(QA[0-9]{2}[A-Z]{4}[A-Z0-9]{21})|(RO[0-9]{2}[A-Z]{4}[A-Z0-9]{16})|(RS[0-9]{20})|(RU[0-9]{31})|(SA[0-9]{4}[A-Z0-9]{18})|(SC[0-9]{2}[A-Z]{4}[0-9]{20}[A-Z]{3})|(SD[0-9]{16})|(SE[0-9]{22})|(SI[0-9]{17})|(SK[0-9]{22})|(SM[0-9]{2}[A-Z]{1}[0-9]{10}[A-Z0-9]{12})|(SO[0-9]{21})|(ST[0-9]{23})|(SV[0-9]{2}[A-Z]{4}[0-9]{20})|(TL[0-9]{21})|(TN[0-9]{22})|(TR[0-9]{8}[A-Z0-9]{16})|(UA[0-9]{8}[A-Z0-9]{19})|(VA[0-9]{20})|(VG[0-9]{2}[A-Z]{4}[0-9]{16})|(XK[0-9]{18})",
"allowPartialMatch": false,
"checksumType": "MOD97",
"matchStrength": 0.9
}
],
"rejectStrength": 0.1
}
},
{
"domain": "IBAN",
"name": "IBAN - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "15",
"typeName": "String" }
]
}
},
{
"domain": "IP ADDRESS",
"name": "IP Address - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(ip_?-? ?addre?s?s?))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "IP ADDRESS",
"name": "IP Address - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "10",
"typeName": "String" }
]
}
},
{
"domain": "IP ADDRESS",
"name": "IP Address - Regex",
"type": "REGEX",
"properties": {
"dataPatterns": [
{
"regex": "(?>((25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\\.){3}(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?))",
"note": "IPv4",
"allowPartialMatch": false,
"matchStrength": 0.9
},
{
"regex": "(?>([A-F0-9]{0,4}:){2,7}[A-F0-9]{0,4})",
"note": "IPv6 standard addresses",
"allowPartialMatch": false,
"matchStrength": 0.9
},
{
"regex": "(?>(([A-F0-9]{0,4}:){2,6})((25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\\.){3}(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?))",
"note": "IPv6 dual addresses",
"allowPartialMatch": false,
"matchStrength": 0.9
}
],
"rejectStrength": 0.1
}
},
{
"domain": "JOB_TITLE",
"name": "Job Title - List",
"type": "LIST",
"properties": {
"valueLists": [
{
"file": "file://job_titles.txt",
"matchStrength": 1.0
}
]
}
},
{
"domain": "JOB_TITLE",
"name": "Job Title - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?>(job)[-_ ]?(title[-_ ])?(name|nm)|job[-_ ]?title|title[-_ ]?(name|nm))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "JOB_TITLE",
"name": "Job Title - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "7",
"typeName": "String" }
]
}
},
{
"domain": "LANGUAGE",
"name": "Language - List",
"type": "LIST",
"properties": {
"valueLists": [
{
"file": "file://languages.txt",
"matchStrength": 1.0
}
]
}
},
{
"domain": "LANGUAGE",
"name": "Language - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?>la?nguage)" }
],
"rejectStrength": 0.0
}
},
{
"domain": "LANGUAGE",
"name": "Language - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "7",
"typeName": "String" }
]
}
},
{
"domain": "LAST_NAME",
"name": "Last Name - List",
"type": "LIST",
"properties": {
"valueLists": [
{
"file": "file://us_last.txt",
"matchStrength": 1.0
},
{
"file": "file://de_last.txt",
"matchStrength": 1.0
},
{
"file": "file://ch_last.txt",
"matchStrength": 1.0
}
]
}
},
{
"domain": "LAST_NAME",
"name": "Last Name - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>((l(as)?t)_?-? ?(na?me?))(_?-?user)?)" },
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(sur) ?_?-? ?(name)?_?-? ?(no|id|str|value|))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "LAST_NAME",
"name": "Last Name - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "10",
"typeName": "String" }
]
}
},
{
"domain": "PLATE_NO",
"name": "License Plate - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(license|li?c)?[-_ ]?plate[-_ ]?(number|num|nbr|no)?)" }
],
"rejectStrength": 0.0
}
},
{
"domain": "PLATE_NO",
"name": "License Plate - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "1",
"typeName": "String" }
]
}
},
{
"domain": "LOCATION_COORDINATES",
"name": "Location Coordinates - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>((map|geographica?l?|areal|location|lat(itude)?|lo?ng(itude)?)[-_ ]?(co-?o?rd(inates?)?)|(latitude|longitude|co-?o?rdinates?)([-_ ]value)?))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "LOCATION_COORDINATES",
"name": "Location Coordinates - Regex",
"type": "REGEX",
"properties": {
"dataPatterns": [
{
"note": "Latitude and Longitude in decimal degrees format",
"regex": "(?!^[+-]?\\d+[. ]\\d+([. ]\\d+)?$)(?:[+-]?([1-8]?[0-9])[. °](\\d{4,8})°?[NS]?,? ?)?[+-]?([1-9]?[0-9]|1[0-7][0-9])[. °](\\d{4,8})°?[NEWS]?",
"allowPartialMatch": false,
"matchStrength": 0.9
},
{
"note": "Latitude or Longitude in decimal degrees format (float value format with low confidence)",
"regex": "(?:[+-]?([1-9]?[0-9]|1[0-7][0-9])[. ]\\d{4,8})",
"allowPartialMatch": false,
"matchStrength": 0.05
},
{
"note": "Latitude & longitude in degrees & minutes format",
"regex": "(?:[+-]?([1-8]?[0-9])[. °]([0-5]?[0-9])\\.(\\d{3,4})(?:'|′)?[NS]?,? ?)?[+-]?([1-9]?[0-9]|1[0-7][0-9])[. °]([0-5]?[0-9])\\.(\\d{3,4})(?:'|′)?[NEWS]?",
"allowPartialMatch": false,
"matchStrength": 0.9
},
{
"note": "Latitude and Longitude in degrees, minutes & seconds format",
"regex": "(?:[+-]?([1-8]?[0-9])[. °]([0-5]?[0-9])(?:'|′)([0-5]?[0-9])(\\.\\d{1,2})?(?:\"|″)[NS]?,? ?)?[+-]?([1-9]?[0-9]|1[0-7][0-9])[. °]([0-5]?[0-9])(?:'|′)([0-5]?[0-9])(\\.\\d{1,2})?(?:\"|″)[NEWS]?",
"allowPartialMatch": false,
"matchStrength": 0.9
}
],
"rejectStrength": 0.1
}
},
{
"domain": "LOCATION_COORDINATES",
"name": "Location Coordinates - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "8",
"typeName": "String" },
{
"minimumLength": "8",
"typeName": "Number" }
]
}
},
{
"domain": "MARITAL_STATUS",
"name": "Marital Status - List",
"type": "LIST",
"properties": {
"valueLists": [
{
"file": "file://marital_status.txt",
"matchStrength": 1.0
}
]
}
},
{
"domain": "MARITAL_STATUS",
"name": "Marital Status - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?>(marital|marriage|married)[-_ ]?sta?tu?s)" }
],
"rejectStrength": 0.0
}
},
{
"domain": "MARITAL_STATUS",
"name": "Marital Status - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "6",
"typeName": "String" }
]
}
},
{
"domain": "MEDICAL_CODES",
"name": "Medical Codes - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?>(med(ica?l)?|(med(ica?l)?|hosp(ita?l)?)[-_ ]?(condition|cond|procedure|proc|bill(ing)?|diag(nosis)?|diagnostic|pcs|cpt))[-_ ]?(code|cd))" },
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?>(^|[-_ ])(icd[-_ ]?(9|10|11)?[-_ ]?(cm|pcs)?|cpt)[-_ ]?(code|cd)?)" }
],
"rejectStrength": 0.0
}
},
{
"domain": "MEDICAL_CODES",
"name": "Medical Codes - Regex",
"type": "REGEX",
"properties": {
"dataPatterns": [
{
"note": "ICD-9 codes: Legacy medical disease and diagnostic codes used world-wide",
"regex": "\\d{3}\\.\\d{0,2}|E\\d{3}\\.?\\d?|V\\d{2}\\.?\\d{0,2}",
"allowPartialMatch": false,
"matchStrength": 0.05
},
{
"note": "ICD-10-CM codes: Newer medical disease and diagnostic codes used world-wide",
"regex": "[A-Z][0-9][0-9AB]\\.?[0-9A-Z]{0,4}",
"allowPartialMatch": false,
"matchStrength": 0.15
},
{
"note": "ICD-10-PCS (Procedure Coding System) codes: Medical inpatient procedure codes used world-wide.",
"regex": "[0-9A-HJ-NP-Z]{7}",
"allowPartialMatch": false,
"matchStrength": 0.05
},
{
"note": "CPT (Current Procedural Terminology) codes (end with alpha regex): Medical procedure codes used mainly in USA and a few handful of countries.",
"regex": "[0-9]{4}[A-Z]",
"allowPartialMatch": false,
"matchStrength": 0.2
},
{
"note": "CPT (Current Procedural Terminology) codes (no alpha regex): Medical procedure codes used mainly in USA and a few handful of countries.",
"regex": "[0-9]{4}[0-9]",
"allowPartialMatch": false,
"matchStrength": 0.05
}
],
"rejectStrength": 0.1
}
},
{
"domain": "MEDICAL_CODES",
"name": "Medical Codes - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "3",
"typeName": "String" },
{
"minimumLength": "3",
"typeName": "Number" }
]
}
},
{
"domain": "MEDICAL_DRUG",
"name": "Medical Drug - List",
"type": "LIST",
"properties": {
"valueLists": [
{
"file": "file://medical_drugs_generic.txt",
"matchStrength": 1.0
},
{
"file": "file://medical_drugs_brand.txt",
"matchStrength": 1.0
}
],
"tokenizeInput": true,
"tokenizationDelimiter": " " }
},
{
"domain": "MEDICAL_DRUG",
"name": "Medical Drug - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?>(med|medical|prescription)?[-_ ]?drug([-_ ]?(prescription)?)(name|nm)?|(medicine?|medication)[-_ ]?(name|nm))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "MEDICAL_DRUG",
"name": "Medical Drug - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "6",
"typeName": "String" }
]
}
},
{
"domain": "NATIONAL_ORIGIN",
"name": "National Origin - List",
"type": "LIST",
"properties": {
"valueLists": [
{
"file": "file://national_origins.txt",
"matchStrength": 1.0
}
]
}
},
{
"domain": "NATIONAL_ORIGIN",
"name": "National Origin - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?>nationa?li?ty|nat(ion)?a?l[-_ ]?(ori?gi?ns?))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "NATIONAL_ORIGIN",
"name": "National Origin - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "2",
"typeName": "String" }
]
}
},
{
"domain": "PO_BOX",
"name": "PO Box - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(p.?o.?_?-? ?box|post_?-? ?office_?-? ?box ?_?-?)(number|num|nbr|no)?)" }
],
"rejectStrength": 0.0
}
},
{
"domain": "PO_BOX",
"name": "PO Box - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "4",
"typeName": "String" },
{
"minimumLength": "4",
"typeName": "Number" }
]
}
},
{
"domain": "PASSWORD",
"name": "Password - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(pass(?!po?rt)) ?_?-??(word)?_?-? ?(word|nbr|no|id|value|))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "PASSWORD",
"name": "Password - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "6",
"typeName": "String" }
]
}
},
{
"domain": "PAYMENT_AMOUNT",
"name": "Payment Amount - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>[-_ ]?(amount|amnt|amt|balance|price)($|[ _-]))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "PAYMENT_AMOUNT",
"name": "Payment Amount - Regex",
"type": "REGEX",
"properties": {
"dataPatterns": [
{
"note": "International currency symbols and abbreviations from: https://wise.com/gb/blog/world-currency-symbols. Regex prefixed with symbol",
"regex": "((EUR|ALL|BYN|BAM|BGN|HRK|CZK|DKK|HRK|GEL|DKK|HUF|ISK|CHF|MDL|MKD|NOK|PLN|RON|RUB|RSD|SEK|CHF|TRY|UAH|GBP|USD|XCD|AWG|ARS|BSD|BBD|BMD|BZD|BOB|BRL|CAD|KYD|CLP|COP|CRC|CUP|ANG|DOP|FKP|GTQ|GYD|HTG|HNL|JMD|MXN|NIO|PAB|PYG|PEN|ANG|SRD|TTD|UYU|VED|AFN|AMD|AZN|BHD|EUR|GEL|IQD|IRR|ILS|JOD|KWD|LBP|ILS|SYP|AED|ILS|OMR|QAR|SAR|YER|XAF|XOF|DZD|AOA|BWP|BIF|CVE|KMF|CDF|DJF|EGP|ERN|ETB|SZL|GMD|GHS|GNF|KES|LSL|LRD|LYD|MGA|MWK|MUR|MRU|MAD|MZN|NAD|NGN|RWF|STN|SCR|SLL|SOS|ZAR|SSP|SDG|TZS|TND|UGX|USD|AUD|BDT|BTN|BND|KHR|CNY|HKD|IDR|INR|JPY|KZT|KGS|LAK|MOP|MYR|MVR|MNT|MMK|NPR|NZD|KPW|PKR|PHP|SGD|KRW|LKR|TWD|TJS|USD|THB|TMT|UZS|VND|BTC|XBT|ETH|LTC|XMR|XRP|USDT)|(€|L|Br|KM|лв|kn|Kč|kn|₾|ft|kr|Íkr|CHF|L|ден|zł|lei|₽|RSD|CHF|₺|₴|£|\\$|ƒ|B\\$|BZ\\$|Bs|R\\$|CA\\$|CI\\$|₡|CUC\\$|ƒ|RD\\$|FK£|Q|G\\$|G|L|J\\$|C\\$|B/\\.|₲|S/\\.|Sr\\$|TT\\$|\\$U|Bs\\.|؋|֏,դր|₼|\\.د.\\ب|€|ლარი|ع.\\د|﷼|₪|ينار|ك|ل.\\ل|₪|£S|AED|₪|ر.\\ع|ر.\\ق|SR|﷼|FCFA|CFA|دج|Kz|P|FBu|CVE|CF|FC|Fdj|E£|Nkf|Br|L|D|GH₵|FG|KSh|L|LD\\$|LD|Ar|K|₨|UM|DH|MT|N\\$|₦|R₣|Db|SR|Le|Sh\\.So\\.|R|SS£|SDG|TSh|د.\\ت|USh|\\$|A\\$|৳|Nu|B\\$|៛|¥|元|\\$|HK\\$|Rp|₹|₸|som|₭|MOP\\$|RM|MRf|₮|K|Rs|\\$|₩|Rs|₱|S\\$|₩|Rs|NT\\$|TJS|US\\$|฿|m|som|₫|₿|Ξ|Ł|ɱ|XRP|₮))\\s?[+-]?(\\d{1,3}(,\\d{3})*|(\\d+))(\\.\\d{2})?",
"allowPartialMatch": false,
"matchStrength": 0.9
},
{
"note": "International currency symbols and abbreviations from: https://wise.com/gb/blog/world-currency-symbols. Regex suffixed with symbol",
"regex": "[+-]?(\\d{1,3}(,\\d{3})*|(\\d+))(\\.\\d{2})?\\s?((EUR|ALL|BYN|BAM|BGN|HRK|CZK|DKK|HRK|GEL|DKK|HUF|ISK|CHF|MDL|MKD|NOK|PLN|RON|RUB|RSD|SEK|CHF|TRY|UAH|GBP|USD|XCD|AWG|ARS|BSD|BBD|BMD|BZD|BOB|BRL|CAD|KYD|CLP|COP|CRC|CUP|ANG|DOP|FKP|GTQ|GYD|HTG|HNL|JMD|MXN|NIO|PAB|PYG|PEN|ANG|SRD|TTD|UYU|VED|AFN|AMD|AZN|BHD|EUR|GEL|IQD|IRR|ILS|JOD|KWD|LBP|ILS|SYP|AED|ILS|OMR|QAR|SAR|YER|XAF|XOF|DZD|AOA|BWP|BIF|CVE|KMF|CDF|DJF|EGP|ERN|ETB|SZL|GMD|GHS|GNF|KES|LSL|LRD|LYD|MGA|MWK|MUR|MRU|MAD|MZN|NAD|NGN|RWF|STN|SCR|SLL|SOS|ZAR|SSP|SDG|TZS|TND|UGX|USD|AUD|BDT|BTN|BND|KHR|CNY|HKD|IDR|INR|JPY|KZT|KGS|LAK|MOP|MYR|MVR|MNT|MMK|NPR|NZD|KPW|PKR|PHP|SGD|KRW|LKR|TWD|TJS|USD|THB|TMT|UZS|VND|BTC|XBT|ETH|LTC|XMR|XRP|USDT)|(€|L|Br|KM|лв|kn|Kč|kn|₾|ft|kr|Íkr|CHF|L|ден|zł|lei|₽|RSD|CHF|₺|₴|£|\\$|ƒ|B\\$|BZ\\$|Bs|R\\$|CA\\$|CI\\$|₡|CUC\\$|ƒ|RD\\$|FK£|Q|G\\$|G|L|J\\$|C\\$|B/\\.|₲|S/\\.|Sr\\$|TT\\$|\\$U|Bs\\.|؋|֏,դր|₼|\\.د.\\ب|€|ლარი|ع.\\د|﷼|₪|ينار|ك|ل.\\ل|₪|£S|AED|₪|ر.\\ع|ر.\\ق|SR|﷼|FCFA|CFA|دج|Kz|P|FBu|CVE|CF|FC|Fdj|E£|Nkf|Br|L|D|GH₵|FG|KSh|L|LD\\$|LD|Ar|K|₨|UM|DH|MT|N\\$|₦|R₣|Db|SR|Le|Sh\\.So\\.|R|SS£|SDG|TSh|د.\\ت|USh|\\$|A\\$|৳|Nu|B\\$|៛|¥|元|\\$|HK\\$|Rp|₹|₸|som|₭|MOP\\$|RM|MRf|₮|K|Rs|\\$|₩|Rs|₱|S\\$|₩|Rs|NT\\$|TJS|US\\$|฿|m|som|₫|₿|Ξ|Ł|ɱ|XRP|₮))",
"allowPartialMatch": false,
"matchStrength": 0.9
},
{
"note": "Currency with Numeric digits and without prefix or suffix currency symbols, but ending in decimals",
"regex": "[+-]?(\\d{1,3}(,\\d{3})*|(\\d+))(\\.\\d{2})",
"allowPartialMatch": false,
"matchStrength": 0.3
}
],
"rejectStrength": 0.1
}
},
{
"domain": "PAYMENT_AMOUNT",
"name": "Payment Amount - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "4",
"typeName": "String" },
{
"minimumLength": "4",
"typeName": "Number" }
]
}
},
{
"domain": "POLITICAL_ORIEN",
"name": "Political Orientation - List",
"type": "LIST",
"properties": {
"valueLists": [
{
"file": "file://political_orientations.txt",
"matchStrength": 1.0
}
],
"rejectStrength": 0.3
}
},
{
"domain": "POLITICAL_ORIEN",
"name": "Political Orientation - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?>politi?ca?l[-_ ]?(orie?nt?(ation)?s?|pref(er)?(ence)?s?|ideolo?gy))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "POLITICAL_ORIEN",
"name": "Political Orientation - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "3",
"typeName": "String" }
]
}
},
{
"domain": "ZIP",
"name": "Postcode - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(zip|post|postal)_?-? ?(co?de?)|(zip))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "ZIP",
"name": "Postcode - Regex",
"type": "REGEX",
"properties": {
"dataPatterns": [
{
"regex": "^[0-9]{5}(?:-[0-9]{4})$",
"matchStrength": 0.7
},
{
"regex": "^[0-9]{5}$",
"matchStrength": 0.2
}
],
"rejectStrength": 0.1
}
},
{
"domain": "ZIP",
"name": "Postcode - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "4",
"typeName": "Number" },
{
"minimumLength": "4",
"typeName": "String" }
]
}
},
{
"domain": "PRECINCT",
"name": "Precinct - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>precinct|prcnct)" }
],
"rejectStrength": 0.0
}
},
{
"domain": "PRECINCT",
"name": "Precinct - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "0",
"typeName": "String" },
{
"minimumLength": "0",
"typeName": "Number" }
]
}
},
{
"domain": "RECORD_NO",
"name": "Record Number - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?>(med(ical)?|rec(ord)?|(med(ical)?)[-_ ]?rec(ord)?)[-_ ]?(num(ber)?|nbr|no|id|key))" },
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?>(med(ical)?[-_ ]?)?health([-_ ]?ca?re?)?([-_ ]?med(ical)?)?([-_ ]?rec(ord)?)?[-_ ]?(num(ber)?|nbr|no|id|key))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "RECORD_NO",
"name": "Record Number - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "5",
"typeName": "String" },
{
"minimumLength": "5",
"typeName": "Number" }
]
}
},
{
"domain": "RELIGIOUS_ORIEN",
"name": "Religious Orientation - List",
"type": "LIST",
"properties": {
"valueLists": [
{
"file": "file://religious_orientations.txt",
"matchStrength": 1.0
}
],
"rejectStrength": 0.3
}
},
{
"domain": "RELIGIOUS_ORIEN",
"name": "Religious Orientation - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?>religion|religio?us[-_ ]?(orie?nt?(ation)?s?|pref(er)?(ence)?s?|practice|faith|type|identity))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "RELIGIOUS_ORIEN",
"name": "Religious Orientation - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "3",
"typeName": "String" }
]
}
},
{
"domain": "ROUTING_NO",
"name": "Routing Number - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?>(^|[-_ ])(routing|aba|ach|rtn|routing[-_ ]?transit)[-_ ]?(number|num|nbr|no))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "ROUTING_NO",
"name": "Routing Number - Regex",
"type": "REGEX",
"properties": {
"dataPatterns": [
{
"regex": "\\d{9}",
"allowPartialMatch": false,
"checksumType": "MOD10_ABA",
"matchStrength": 0.9
}
],
"rejectStrength": 0.5
}
},
{
"domain": "ROUTING_NO",
"name": "Routing Number - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "9",
"typeName": "String" },
{
"minimumLength": "9",
"typeName": "Number" }
]
}
},
{
"domain": "SCHOOL_NM",
"name": "School Name - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>school_?-?na?me?)" }
],
"rejectStrength": 0.0
}
},
{
"domain": "SCHOOL_NM",
"name": "School Name - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "20",
"typeName": "String" }
]
}
},
{
"domain": "SECURITY_CODE",
"name": "Security Code - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>se?cu?r(i?ty?)?_?co?de?)" }
],
"rejectStrength": 0.0
}
},
{
"domain": "SECURITY_CODE",
"name": "Security Code - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "5",
"typeName": "String" },
{
"minimumLength": "5",
"typeName": "Number" },
{
"minimumLength": "0",
"typeName": "Binary" }
]
}
},
{
"domain": "SERIAL_NO",
"name": "Serial Number - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(ser(ial)?)_?-? ?(number|num|nbr|no))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "SERIAL_NO",
"name": "Serial Number - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "0",
"typeName": "Number" },
{
"minimumLength": "0",
"typeName": "String" }
]
}
},
{
"domain": "SEXUAL_ORIENTATION",
"name": "Sexual Orientation - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(sexuali?ty|sexual[-_ ]?(orie?nt?(ation)?s?)?(pref(er)?(ence)?s?)?[-_ ]?(type|identity)?))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "SEXUAL_ORIENTATION",
"name": "Sexual Orientation - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "5",
"typeName": "String" }
]
}
},
{
"domain": "SEXUAL_ORIENTATION",
"name": "Sexual Orientation - List",
"type": "LIST",
"properties": {
"valueLists": [
{
"file": "file://sexual_orientations.txt",
"matchStrength": 0.9
}
],
"rejectStrength": 0.1
}
},
{
"domain": "SIGNATURE",
"name": "Signature - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(signature))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "SIGNATURE",
"name": "Signature - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "0",
"typeName": "String" },
{
"minimumLength": "0",
"typeName": "Binary" }
]
}
},
{
"domain": "SSN",
"name": "Social Security Number - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(ssn$|social_?-? ?security_?-? ?(number|num|nbr|no|code|id)))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "SSN",
"name": "Social Security Number - Regex",
"type": "REGEX",
"properties": {
"dataPatterns": [
{
"regex": "(?!000)(?!666)[0-8]\\d{2}([- ])(?!00)\\d{2}\\1(?!0000)\\d{4}",
"allowPartialMatch": false,
"matchStrength": 0.7
},
{
"regex": "[9]\\d{2}([- ])(?!(6[6-9]))(?!89)([5-9][0-9])\\1\\d{4}",
"allowPartialMatch": false,
"matchStrength": 0.7
},
{
"regex": "(?!000)(?!666)[0-8]\\d{2}(?!00)\\d{2}(?!0000)\\d{4}",
"allowPartialMatch": false,
"matchStrength": 0.1
},
{
"regex": "[9]\\d{2}(?!(6[6-9]))(?!89)([5-9][0-9])\\d{4}",
"allowPartialMatch": false,
"matchStrength": 0.1
}
],
"rejectStrength": 0.1
}
},
{
"domain": "SSN",
"name": "Social Security Number - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "9",
"typeName": "Number" },
{
"minimumLength": "9",
"typeName": "String" }
]
}
},
{
"domain": "SWIFT_CODE",
"name": "Swift Code - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?>(swift|(^|[-_ ])bic)[-_ ]?(number|num|nbr|no|code|cd|id)|bank[-_ ]?(identification|identifier)[-_ ]?(code|cd))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "SWIFT_CODE",
"name": "Swift Code - Regex",
"type": "REGEX",
"properties": {
"dataPatterns": [
{
"note": "Swift-code format: 8-11 digit code with country-code at 5,6 positions.",
"regex": "[A-Z]{4}-?[AD|AE|AF|AG|AI|AL|AM|AO|AQ|AR|AS|AT|AU|AW|AX|AZ|BA|BB|BD|BE|BF|BG|BH|BI|BJ|BL|BM|BN|BO|BQ|BR|BS|BT|BV|BW|BY|BZ|CA|CC|CD|CF|CG|CH|CI|CK|CL|CM|CN|CO|CR|CU|CV|CW|CX|CY|CZ|DE|DJ|DK|DM|DO|DZ|EC|EE|EG|EH|ER|ES|ET|FI|FJ|FK|FM|FO|FR|GA|GB|GD|GE|GF|GG|GH|GI|GL|GM|GN|GP|GQ|GR|GS|GT|GU|GW|GY|HK|HM|HN|HR|HT|HU|ID|IE|IL|IM|IN|IO|IQ|IR|IS|IT|JE|JM|JO|JP|KE|KG|KH|KI|KM|KN|KP|KR|KW|KY|KZ|LA|LB|LC|LI|LK|LR|LS|LT|LU|LV|LY|MA|MC|MD|ME|MF|MG|MH|MK|ML|MM|MN|MO|MP|MQ|MR|MS|MT|MU|MV|MW|MX|MY|MZ|NA|NC|NE|NF|NG|NI|NL|NO|NP|NR|NU|NZ|OM|PA|PE|PF|PG|PH|PK|PL|PM|PN|PR|PS|PT|PW|PY|QA|RE|RO|RS|RU|RW|SA|SB|SC|SD|SE|SG|SH|SI|SJ|SK|SL|SM|SN|SO|SR|SS|ST|SV|SX|SY|SZ|TC|TD|TF|TG|TH|TJ|TK|TL|TM|TN|TO|TR|TT|TV|TW|TZ|UA|UG|UM|US|UY|UZ|VA|VC|VE|VG|VI|VN|VU|WF|WS|YE|YT|ZA|ZM|ZW]-?[A-Z0-9]{2}(-[A-Z0-9]{3})?",
"allowPartialMatch": false,
"matchStrength": 0.2
}
],
"rejectStrength": 0.3
}
},
{
"domain": "SWIFT_CODE",
"name": "Swift Code - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "8",
"typeName": "String" }
]
}
},
{
"domain": "TAX_ID",
"name": "Tax ID - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(tax)_?-? ?(id(ent)?)_?-? ?((co?de?)|(number|num|nbr|no))?)" },
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>tin$)" }
],
"rejectStrength": 0.0
}
},
{
"domain": "TAX_ID",
"name": "Tax ID - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "6",
"typeName": "String" },
{
"minimumLength": "6",
"typeName": "Number" }
]
}
},
{
"domain": "TELEPHONE_NO",
"name": "Telephone Number - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(phone|contact|^tel($|[ _-])|fax|mobile|telephone)_?-? ?)(number|num|nbr|no)?" }
],
"rejectStrength": 0.0
}
},
{
"domain": "TELEPHONE_NO",
"name": "Telephone Number - Regex",
"type": "REGEX",
"properties": {
"dataPatterns": [
{
"regex": "((\\+?1|001)[-. ]?)?(\\([0-9]{3}\\)[-. ]?|[0-9]{3}[-. ])[0-9]{3}[-. ][0-9]{4}",
"allowPartialMatch": false,
"matchStrength": 0.7
},
{
"regex": "\\+(?:[0-9][ .()\\/-]?){6,14}[0-9]",
"allowPartialMatch": false,
"matchStrength": 0.5
},
{
"regex": "[0-9]{5,17}",
"allowPartialMatch": false,
"dataCleanRegex": "[+ -.()\\/]",
"matchStrength": 0.05
}
],
"rejectStrength": 0.1
}
},
{
"domain": "TELEPHONE_NO",
"name": "Telephone Number - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "7",
"typeName": "Number" },
{
"minimumLength": "7",
"typeName": "String" }
]
}
},
{
"domain": "UNION_MEMBERSHIP_NO",
"name": "Union Membership Number - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(union)[_ -]?)(membe?r(ship)?)?[_ -]?(number|num|nbr|no|id)(?!\\w*(name|group|grp))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "UNION_MEMBERSHIP_NO",
"name": "Union Membership Number - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "3",
"typeName": "String" },
{
"minimumLength": "3",
"typeName": "Number" }
]
}
},
{
"domain": "US_PASSPORT",
"name": "US Passport - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(passport)[_ -]?)(number|num|nbr|no)?" }
],
"rejectStrength": 0.0
}
},
{
"domain": "US_PASSPORT",
"name": "US Passport - Regex",
"type": "REGEX",
"properties": {
"dataPatterns": [
{
"regex": "[A-Z]\\d{8}",
"allowPartialMatch": false,
"matchStrength": 0.2
},
{
"regex": "(?!0{9})\\d{9}",
"allowPartialMatch": false,
"matchStrength": 0.1
}
],
"rejectStrength": 0.1
}
},
{
"domain": "US_PASSPORT",
"name": "US Passport - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "9",
"typeName": "String" },
{
"minimumLength": "9",
"typeName": "Number" }
]
}
},
{
"domain": "US_STATE",
"name": "US State - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)state",
"allowPartialMatch": false
}
],
"rejectStrength": 0.0
}
},
{
"domain": "US_STATE",
"name": "US State - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "14",
"typeName": "String" }
]
}
},
{
"domain": "US_STATE",
"name": "US State - List",
"type": "LIST",
"properties": {
"valueLists": [
{
"file": "file://us_states_full.txt",
"matchStrength": 1.0
}
],
"rejectStrength": 0.1
}
},
{
"domain": "USPS_STATE_CODE",
"name": "USPS State Code - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)state[ _-]?(cd|code|abbrev)?",
"allowPartialMatch": false
}
],
"rejectStrength": 0.0
}
},
{
"domain": "USPS_STATE_CODE",
"name": "USPS State Code - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "2",
"typeName": "String" }
]
}
},
{
"domain": "USPS_STATE_CODE",
"name": "USPS State Code - List",
"type": "LIST",
"properties": {
"valueLists": [
{
"file": "file://us_states.txt",
"matchStrength": 1.0
}
],
"rejectStrength": 0.5
}
},
{
"domain": "VIN_NO",
"name": "Vehicle Identification Number - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(^vin$|vehicle_?-? ?id(entification)?_?-? ?(number|num|nbr|no)))" },
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(vehicle))" }
],
"rejectStrength": 0.0
}
},
{
"domain": "VIN_NO",
"name": "Vehicle Identification Number - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "10",
"typeName": "String" }
]
}
},
{
"domain": "WEB",
"name": "Web URL - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(^url_?-? ?|web_? ?)(addr?e?s?s?)?)" }
],
"rejectStrength": 0.0
}
},
{
"domain": "WEB",
"name": "Web URL - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "10",
"typeName": "String" }
]
}
},
{
"domain": "USER_ID",
"name": "User ID - Path",
"type": "PATH",
"properties": {
"paths": [
{
"matchType": "REGEX",
"matchStrength": 0.67,
"fieldValue": "(?i)(?>(use?r)([-_ ])?(na?me?|id)?)$" }
],
"rejectStrength": 0.0
}
},
{
"domain": "USER_ID",
"name": "User ID - Type",
"type": "TYPE",
"properties": {
"allowedTypes": [
{
"minimumLength": "8",
"typeName": "String" },
{
"minimumLength": "8",
"typeName": "Number" }
]
}
}
]
The ASDD Standard profile set can identify the following PII:
ASDD Standard identifiable PII
Account Number Address Line 1 Address Line 2 Age Bank Account Number Beneficicary Number Biometric Blood Type Certificate Number City Company Name Country County Credit Card Number Customer Number Date of Birth Department Drivers License Email Address Ethnicity First Name Text Full Name Gender HIPAA Date House Number IBAN IP Address Job Title Language Last Name License Plate Location Coordinate Marital Status Medical Codes Medical Drug National Origin PO Box Password Payment Amount Political Orientation Postcode Precinct Record Number Religious Orientation Routing Number School Name Security Code Serial Number Sexual Orientation Signature Social Security Number Swift Code Tax ID Telephone Number Union Membership Number US Passport US State USPS State Code Vehicle Identification Number Web URL User ID