Token Functions

reference

Tokenization is the process of breaking a stream of text up into words, phrases, symbols, or other meaningful elements called tokens. The list of tokens becomes input for further processing such as parsing or text mining. Token functions are not limited to string input since they work with generic JSON objects and documents.

If any arguments to any of the following functions are MISSING then the result is also MISSING (i.e. no result is returned). Similarly, if any of the arguments passed to the functions are NULL or are of the wrong type (e.g. an integer instead of a string), then NULL is returned as the result.

CONTAINS_TOKEN(input_obj, token_expr [, options ])

Description

Checks whether or not the specified search token token_expr is a sub-string of the input object input_obj.

Arguments

input_obj

Any JSON object, or any valid expression which evaluates to a string, to search within.

token_expr

A token string, or any valid expression which evaluates to a string, that is being searched for.

options

An optional JSON object to control tokenization. Within options:

names: A boolean to include object names (default: true)
case: Either "lower" or "upper" for case folding (default: no change to the original text)
specials: A boolean to include strings with special characters, such as email addresses and URLs (default: false)
split: A boolean to split string values into words (default: true)
trim: A boolean to trim spaces around unsplit string values (default: true)

Return Value

A boolean, representing whether the search expression exists within the input object.

This returns true if the sub-string exists within the input string, otherwise false is returned.

Examples

To try the examples in this section, set the query context to the inventory scope in the travel sample dataset. For more information, see Query Context.

Finding hotels with the word "Inn" within their name

SELECT name
FROM hotel
WHERE CONTAINS_TOKEN(name, "Inn",{"specials":true})
LIMIT 4;

[
    {
        "name": "Sportsman Inn"
    },
    {
        "name": "Keefer's Inn"
    },
    {
        "name": "Quality Inn King City Hotel"
    },
    {
        "name": "Premier Inn, Albert Dock"
    }
]

CONTAINS_TOKEN_LIKE(input_obj, token_expr [, options ])

Description

Checks whether or not the specified search token token_expr is a sub-string of the input object input_obj.

Arguments

input_obj

Any JSON object, or any valid expression which evaluates to a string, to search within.

token_expr

A token string, or any valid expression which evaluates to a string, that is being searched for.

options

An optional JSON object to control tokenization. Within options:

names: A boolean to include object names (default: true)
case: Either "lower" or "upper" for case folding (default: no change to the original text)
specials: A boolean to include strings with special characters, such as email addresses and URLs (default: false)
split: A boolean to split string values into words (default: true)
trim: A boolean to trim spaces around unsplit string values (default: true)

Return Value

A boolean, representing whether the search expression exists within the input object.

This returns true if the sub-string exists within the input string, otherwise false is returned.

Examples

To try the examples in this section, set the query context to the inventory scope in the travel sample dataset. For more information, see Query Context.

Finding email addresses of UK hotels

SELECT email
FROM hotel
WHERE CONTAINS_TOKEN_LIKE(email, "%uk",{"specials":true})
LIMIT 4;

[
    {
        "email": "glencoe@syha.org.uk"
    },
    {
        "email": "owner@hillhousellanrhidian.co.uk"
    },
    {
        "email": "julia@number38thegower.co.uk"
    },
    {
        "email": "stay@holiday-harlech.co.uk"
    }
]

CONTAINS_TOKEN_REGEXP(input_obj, token_expr [, options ])

Description

Checks whether or not the specified search token token_expr is a sub-string of the input object input_obj.

Arguments

input_obj

Any JSON object, or any valid expression which evaluates to a string, to search within.

token_expr

A token string, or any valid expression which evaluates to a string, that is being searched for.

options

An optional JSON object to control tokenization. Within options:

names: A boolean to include object names (default: true)
case: Either "lower" or "upper" for case folding (default: no change to the original text)
specials: A boolean to include strings with special characters, such as email addresses and URLs (default: false)
split: A boolean to split string values into words (default: true)
trim: A boolean to trim spaces around unsplit string values (default: true)

Return Value

A boolean, representing whether the search expression exists within the input object.

This returns true if the sub-string exists within the input string, otherwise false is returned.

Examples

To try the examples in this section, set the query context to the inventory scope in the travel sample dataset. For more information, see Query Context.

Finding hotels with the word "Inn" within their name

SELECT name
FROM hotel
WHERE CONTAINS_TOKEN_REGEXP(name, "In+.*",{"specials":true})
LIMIT 4;

[
    {
        "name": "Sportsman Inn"
    },
    {
        "name": "Inveraray Youth Hostel"
    },
    {
        "name": "Inverness Youth Hostel"
    },
    {
        "name": "Indian Cove Campground"
    }
]

HAS_TOKEN(input_obj, token_expr [, options ])

Alias for CONTAINS_TOKEN().

TOKENS(in_str, opt)

Description

This function tokenizes (i.e. breaks up into meaningful segments) the given input string based on specified delimiters, and other options. It recursively enumerates all tokens in a JSON value and returns an array of values (JSON atomic values) as the result.

Arguments

in_str

A valid JSON object, this can be anything: constant literal, simple JSON value, JSON key name or the whole document itself.

The following table lists the rules for each JSON type:

JSON Type	Return Value
MISSING	[]
NULL	[NULL]
false	[false]
true	[true]
number	[number]
string	SPLIT(string)
array	FLATTEN(TOKENS(element) for each element in array (Concatenation of element tokens)
object	For each name-value pair, name+TOKENS(value)

JSON Type

Return Value

MISSING

[]

NULL

[NULL]

false

[false]

true

[true]

number

[number]

string

SPLIT(string)

array

FLATTEN(TOKENS(element) for each element in array

(Concatenation of element tokens)

object

For each name-value pair, name+TOKENS(value)

opt

A JSON object indicating the options passed to the TOKENS() function. Options can take the following options, and each invocation of TOKENS() can choose one or more of the options:

{"name": true}

Optional. Valid values are true or false. By default, this is set to true and TOKENS() will include field names. You can choose to not include field names by setting this option to false.

{"case":"lower"}

Optional. Valid values are lower or upper. Default is neither, as in it returns the case of the original data. Use this option to specify the case sensitivity.

{"specials": true}

Optional. Use this option to preserve strings with specials characters, such as email addresses, URLs, and hyphenated phone numbers. The default value is false.

The specials options preserves special characters except at the end of a word.

Return Value

An array of strings containing all of the tokens obtained from the input string.

Examples

By default, for speed, the results are randomly ordered. To make the difference more clear between the first two example queries, the ARRAY_SORT() function is used.

List the tokens of an array where specials is FALSE

SELECT ARRAY_SORT(
  TOKENS( ['jim@example.com, kim@example.com, http://example.com/, 408-555-1212'],
          {'specials': false} ));

[
  {
    "$1": [
      "1212",
      "408",
      "555",
      "abc",
      "com",
      "http",
      "jim",
      "kim"
    ]
  }
]

List the tokens of an array where specials is TRUE

SELECT ARRAY_SORT(
  TOKENS( ['jim@example.com, kim@example.com, http://example.com/, 408-555-1212'],
          {'specials': true} ));

[
  {
    "$1": [
      "1212",
      "408",
      "408-555-1212",
      "555",
      "abc",
      "com",
      "http",
      "http://example.com",
      "jim",
      "jim@example.com",
      "kim",
      "kim@example.com"
    ]
  }
]

Convert all of the URL data into UPPER case and adds the full URL to the delimited words

For this example, set the query context to the inventory scope in the travel sample dataset. For more information, see Query Context.

SELECT ARRAY_SORT( TOKENS(url) ) AS defaulttoken,
       ARRAY_SORT( TOKENS(url, {"specials":true, "case":"UPPER"}) ) AS specialtoken
FROM hotel
LIMIT 1;

[
  {
    "defaulttoken": [
      "http",
      "org",
      "uk",
      "www",
      "yha"
    ],
    "specialtoken": [
      "HTTP",
      "HTTP://WWW.YHA.ORG.UK",
      "ORG",
      "UK",
      "WWW",
      "YHA"
    ]
  }
]

You can also use {"case":"lower"} or {"case":"upper"} to have case sensitive search. Index creation and querying can use this and other parameters in combination. These parameters should be passed within the query predicates as well. The parameters and values must match exactly for SQL++ to pick up and use the index correctly.

Create an index with case and use it your application

For this example, set the query context to the inventory scope in the travel sample dataset. For more information, see Query Context.

CREATE INDEX idx_url_upper_special ON hotel(
    DISTINCT ARRAY v FOR v IN
        TOKENS(url, {"specials":true, "case":"UPPER"})
    END );

SELECT name, address, url
FROM hotel
WHERE ANY v IN TOKENS(url, {"specials":true, "case":"UPPER"})
      SATISFIES v = "HTTP://WWW.YHA.ORG.UK"
      END;

{
    "results": [
        {
            "address": "Capstone Road, ME7 3JE",
            "name": "Medway Youth Hostel",
            "url": "http://www.yha.org.uk"
        }
    ]
}