Splitting texts inputs

You can split user inputs based on regular expressions, it will result on a multi value search.

Split user input to make a multi value search from fulltext inputs :

Search for ids from fulltext inputs :

This behavior is considered as bad practice. It is documented and present in basic config files for legacy purpose. Instead we recommend to create a specific filter to search on specific fields.

How to tune this behavior :

// /dam/explore/full-text-search.json

{
    "regexId": "^#\\d+$",
    "pattern": {
      "regexps": ["\\d+"],
      "separators": ",; ",
    }
}

By default splitted terms must all be present to match a result.

Meaning that searching for teapot copie will match teapot SOMETHING copie but not teapot or copie.

If you want a more google like experience, where your search for teapot copie will match teapot SOMETHING copie AND teapot AND copie.

You have to add an extra config key :

// /dam/explore/full-text-search.json

{
    "regexId": "^#\\d+$",
    "forceModeShouldOnPattern": true,
    "pattern": {
      "regexps": ["\\d+"],
      "separators": ",; ",
    }
}

Split user input to make a multi value search from advanced search text inputs :

How to tune this behavior :

// /dam/explore/advanced-filters/default-cursor-config/default-resource-config.json

[
  {
    "field": "name",
    "i18nForEmpty": "advancedsearch.field.name.empty",
    "i18nForNotEmpty": "advancedsearch.field.name.notempty",
    "operators": [
      "contains",
      "startswith",
      "endswith",
      "notcontains",
      "empty",
      "notempty"
    ],
    /**
     * Patterns are regepx that you want use on the field to split the user input
     * Example : for user input '1111111111111 2222222222222'
     * The corresponding filter will be :
     * Names contains '1111111111111' OR '2222222222222'
     * 
     * The values contained in the "regexps" array correspond to one or several
     * "named-patterns" (defined in : _portal/common/named-patterns)
     * You can use your own regular expression.
     * 
    */
    "pattern": {
      "regexps": ["ean"]
    }
},
{
    "field": "description",
    "i18nForEmpty": "advancedsearch.field.caption.empty",
    "i18nForNotEmpty": "advancedsearch.field.caption.notempty",
    "operators": [
      "contains",
      "startswith",
      "endswith",
      "notcontains",
      "empty",
      "notempty"
    ],
    /**
     * You can have more control on the provided pattern by adding values to the pattern entry.
     * 
     * Note that you can add directly a custom regex ("\\d{2}").
     * If you use it, you must escape the backslash !
     *
     * "flag" let you control how the pattern should behave.
     * See : https://javascript.info/regexp-introduction#flags
     *
     * "separators" provides you a way to control which characters must be considered as separators
     *
     * Here : for user input '11-22 vv-666-yy'
     * The corresponding filter will be :
     * Description contains '11' OR '22' OR 'vv-666-yy'
     *
     * Note that the unmatching parts of the user input will be put together in another OR
     *
     * Example : for user input '11-22-777 vv-666-yy-999'
     * The corresponding filter will be :
     * Description contains '11' OR '22' OR 'vv-666-yy' OR '777 999'
    */
    "pattern": {
      "regexps": [ "immatriculation", "\\d{2}" ],
      "separators": "-_",
      "flag": "ig"
    }
  }
]