Syntax for changing schema settings

Use shortcuts field, fieldType, and copyField to:

• Add or drop table columns from the index, for example:

  ALTER SEARCH INDEX SCHEMA ON demo.health_data ADD field gender;

  If the field name matches a column name the field definition is automatically added to the pending schema.

• Identify the element (field, fieldType, and copyField) and then change the setting using an element path or JSON definition:

  ALTER SEARCH INDEX SCHEMA ON wiki.solr ADD copyField[@source='title', @dest='catch_all'];

The element path uniquely describes the setting in the schema XML. Enclose attributes after an element in brackets; to define multiple attributes use a comma-separated list. When adding an element, include all of the attributes.

top_level_element_name.child_element_name[@attribute_name='value', ...]

For example to add a the Text field type definition:

types.fieldType[ @name='TextField_intl' , @class='org.apache.solr.schema.TextField' ]

You can use element_path to change a sub-element in the schema. For example, to change the ASCIIFoldingFilterFactory in the search analyzer to a ClassicFilterFactory:

ALTER SEARCH INDEX SCHEMA ON demo.users
  SET types.fieldType[@name='TextField_intl']
           .analyzer[@type='search']
           .filter[@class='solr.ASCIIFoldingFilterFactory']@class=
           'solr.ClassicFilterFactory';

Changes the fieldType to:

<fieldType class="org.apache.solr.schema.TextField" name="TextField_intl">
  ...
  <analyzer type="search">
    <filter class="solr.LowerCaseFilterFactory"/>
    <filter class="solr.ClassicFilterFactory"/>
    <tokenizer class="solr.StandardTokenizerFactory"/>
  </analyzer>
</fieldType>

This JSON snippet is translated into XML elements and attributes:

• JSON pair translates to XML attribute.
• JSON object translates to XML element.

The JSON is translated into these XML attributes:

$${
    "analyzer": [
      {
        "type": "index",
        "tokenizer": { "class": "solr.StandardTokenizerFactory" },
        "filter": [
          { "class": "solr.LowerCaseFilterFactory" },
          { "class": "solr.ASCIIFoldingFilterFactory" }
        ]
      },
      {
        "type": "search",
        "tokenizer": { "class": "solr.StandardTokenizerFactory" },
        "filter": [
          { "class": "solr.LowerCaseFilterFactory" },
          { "class": "solr.ASCIIFoldingFilterFactory" }
        ]
      }
    ]
}$$

The JSON is translated into these XML elements:

"analyzer": [				
{ "type": "index", <analyzer type="index">
  "tokenizer":                         
  { "class": "solr.Standard..." }, <tokenizer class="solr.Standard..."/>
    "filter": [				  
    { "class": "solr.LowerCase..." }, <filter class="solr.LowerCase..."/>
    { "class": "solr.ASCII..." } <filter class="solr.ASCII..."/>

The CQL command syntax to remove the second filter on the search phase analyzer:

ALTER SEARCH INDEX SCHEMA ON demo.users
  DROP types.fieldType[@name='TextField_intl']
            .analyzer[@type='search']
            .filter[@class='solr.ClassicFilterFactory'];

Changes the fieldType to:

<fieldType class="org.apache.solr.schema.TextField" name="TextField_intl">
  ...
<analyzer type="search">
    <filter class="solr.LowerCaseFilterFactory"/>
    <tokenizer class="solr.StandardTokenizerFactory"/>
  </analyzer>
</fieldType>

The schema displays in XML format. Use element paths to define and identify elements and attributes:

<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<schema name="autoSolrSchema" version="1.5">
  <types>
    <fieldType class="class_path" 
               name="fieldtype_name" >
      <analyzer>
        <tokenizer class="class_path"/>
        <filter class="class_path"/>
      </analyzer>
    </fieldType>
  </types>
  <fields>
    <field attribute_name="value"
          docValues="true|false" 
          indexed="true|false" 
          multiValued="true|false" 
          name="column_name" 
          stored="true|false" 
          type="fieldtype_name"/>
    
    <copyField 
               source="field_name" 
               dest="field_name" />
  </fields>
  <uniqueKey>pk_column_list</uniqueKey>
</schema>

ADD FIELD column_name

Adds a column from the CQL table to the pending search index schema using the default mapping.

ALTER SEARCH INDEX SCHEMA ON demo.health_data ADD fields.field fips;

Note: Adding the leading element fields. in ADD fields.field fieldname is optional and provides only cosmetic structure.

ADD fields.field[@attribute_name='value', ...]

Adds a new field to the pending schema and manually set the attributes. For example, to add a column from the table to the index and set the field type to string.

ALTER SEARCH INDEX SCHEMA ON demo.health_data 
ADD fields.field[@name='fish', @type='StrField', @indexed='true'];

ADD copyField[@attribute_name='value', ...]

Copy the value of the source field to a new field. For example, as a workaround to the rule that you cannot use LowerCaseStrField on a primary key column, you can use copyField to copy the primary key field data to a new field defined as type LowerCaseStrField.

ALTER SEARCH INDEX SCHEMA ON <table> ADD lowerCaseString key_column_copy;

ALTER SEARCH INDEX SCHEMA ON <table> ADD copyField[@source='key_column', @dest='key_column_copy'];

DROP field field_name

Removes a field from the pending search index schema.

ALTER SEARCH INDEX SCHEMA ON demo.health_data 
DROP field fips;

SET fields.field[@name='field_name']@attribute_name='value'

Changes the field identified by the attribute in brackets by adding or replacing the attribute_to_change.

Field attributes:

• name: Matches a CQL table column name or the name of a copyField destination.
• type: Name of a defined fieldType.
• indexed: True indicates that the field is indexed. By default, only the fields that are included in the index on creation are displayed.
  Note: Primary key columns must be indexed (indexed="true").
• docValues: Creates a forward index on the field values.
• multiValued: Contains more than one value, such as a set, map, list column, or the destination of multiple copyField definitions.

Restriction: The stored field attribute is not supported.

ALTER SEARCH INDEX SCHEMA ON demo.health_data 
SET fields.field[@name='gender_s']@multiValued='true';

ADD types.fieldType[@attribute_name='value', …] WITH $$ { json_map } $$

Adds a field type definition to the schema for analyzing, tokenizing, and filtering fields in the index.

ADD types.fieldType[@name='TrieIntField', @class='solr.TrieIntField']

Note: Optionally add the leading element fields. in ADD and SET fields.field field_name to follow a naming convention and provide structure.