Migrating to the Amazon CloudSearch 2013-01-01 API

Analysis Schemes—you configure analysis schemes to specify language-specific text processing options for text and text-array fields. Amazon CloudSearch now supports 33 languages, as well as an option for multi-language fields. For more information, see Configuring Analysis Schemes. For the complete list of supported languages, see Supported Languages.

Availability Options—you can enable the Multi-AZ option to expand a domain into a second availability zone to ensure availability in the event of a service disruption. For more information, see Configuring Availability Options.

Scaling Options—you can set the desired instance type and desired replication count to increase upload or search capacity, speed up search requests, and improve fault tolerance. For more information, see Configuring Scaling Options in Amazon CloudSearch.

Suggesters—you can configure suggesters to implement autocomplete functionality. For more information, see Configuring Suggesters for Amazon CloudSearch.

Analysis Scheme—you configure language-specific text-processing on a per field basis by specifying an analysis scheme for each text and text-array field. For more information, see Configuring Analysis Schemes.

Field Types—Amazon CloudSearch now supports 11 field types:

Highlight—when you enable the highlight option for a field, you can retrieve excerpts that show where the search terms occur within that field. For more information, see Highlighting Search Hits in Amazon CloudSearch.

Source—you can specify a source for a field to copy data from one field to another, enabling you to use the same source data in different ways by configuring different options for the fields.

By default, when you add a field, all options valid for that field type are enabled. While this is useful for development and testing, disabling options you don't need can reduce the size of your index and improve performance.

You must use the separate array type fields for multi-valued fields.

Only single-value fields can be sort enabled.

Only text and text-array fields can be highlight enabled.

All fields except text and text-array fields can be facet enabled.

Literal fields are now case-sensitive.

You no longer have to store floating point values as integers—use a double field.

You can store locations using the new latlon field type. For more information, see location-based searching and sorting.

An int field is a 64-bit signed integer.

Instead of configuring a default search field, you can specify which fields to search with the q.options parameter in your search requests. The q.options parameter also enables you to specify weights for each of the fields.

When sorting and configuring expressions, you reference the default relevance score with the name _score. Due to changes in the relevance algorithm, the calculated scores will be different than they were under the 2011-02-01 API. For more information, see Configuring Expressions.

Expressions now support the logn, atan2, and haversin functions as well as the _score (text relevance score) and _time (epoch time) variables. If you store locations in latlon fields, you can reference the latitude and longitude values as FIELD.latitude and FIELD.longitude. You can also reference both int and double fields in expressions. The following functions are no longer supported: cs.text_relevance, erf, lgamma, rand, and time. For more information, see Configuring Expressions.

DefineAnalysisScheme

DefineExpression

DefineSuggester

DeleteAnalysisScheme

DeleteExpression

DeleteSuggester

DexcribeAnalysisSchemes

DescribeAvailabilityOptions

DescribeExpressions

DescribeScalingParameters

DescribeSuggesters

ListDomainNames

UpdateAvailabilityOptions

UpdateScalingParameters

DefineRankExpression

DescribeRankExpression

DeleteRankExpression

DescribeDefaultSearchField

DescribeStemmingOptions

DescribeStopwordOptions

DescribeSynonymOptions

UpdateDefaultSearchField

UpdateStemmingOptions

UpdateStopwordOptions

UpdateSynonymOptions

Omit the version and lang attributes from your document batches.

Make sure all of the document fields correspond to index fields configured for your domain. Unrecognized fields are no longer ignored, they will generate an error.

Post the document batches to your 2013-01-01 domain's doc endpoint. Note that you must specify the 2013-01-01 API version. For example, the following request posts the batch contained in data1.json to the doc-movies-123456789012.us-east-1.cloudsearch.amazonaws.com endpoint.

curl -X POST --upload-file data1.json doc-movies-123456789012.us-east-1.
cloudsearch.amazonaws.com/2013-01-01/documents/batch --header "Content-Type:
application/json"

Use the 2013-01-01 API version in all requests.

Use the q parameter to specify search criteria for all requests. The bq parameter is no longer supported. To use the structured (Boolean) search syntax, specify q.parser=structured in the request.

Parameters cannot be repeated in a search request.

The wildcard character (*) is only supported when using the simple query parser. Use the prefix operator to perform prefix matching with the structured query parser. For example, q=(prefix 'oce')&q.parser=structured.

Use the field name _id to reference the document ID field in a search request. The docid field name is no longer supported.

Use the range operator search a field for a value within the specified range. The filter operator is no longer supported.

Use the new range syntax to search for ranges of values, including dates and locations stored in latlon fields. The double dot (..) notation is no longer supported. Separate the upper and lower bounds with a comma (,), and enclose the range in brackets or braces. A square bracket ([,]) indicates that the bound is included, a curly brace ({,}) excludes the bound. For example, year:2008..2011 is now expressed as year:[2008,2011]. An open ended range such as year:..2011 is now expressed as year:{,2011].

Use the term operator to search a field for a particular value. The field operator is no longer supported.

Use the q.options parameter to specify field weights. The cs.text_relevance function is no longer supported. For example, q.options={fields:['title^2','plot^0.5']}.

Use the fq parameter to filter results without affecting how the matching documents are scored and sorted.

Use a dot (.) as a separator rather than a hyphen (-) in the prefix parameters: expr.NAME, facet.FIELD, highlight.FIELD.

Use the facet.FIELD parameter to specify all facet options. The facet-FIELD-top-N, facet-FIELD-sort, and facet-FIELD-constraints parameters are no longer supported.

Use the sort parameter to specify the fields or expressions you want to use for sorting. You must explicitly specify the sort direction in the sort parameter. For example, sort=rank asc, date desc. The rank parameter is no longer supported.

Use expr.NAME to define an expression in a search request. The rank-RANKNAME parameter is no longer supported.

Use format=xml to get results as XML. The result-type parameter is no longer supported.

Term boosting—use the boost option in a structured query to increase the importance of one part of the query relative to the other parts. For more information, see Constructing Compound Queries.

Sloppy phrase search—use the near operator in a structured query to search a text or text-array field for multiple terms and find documents that contain the terms within the specified distance of one another. You can also perform a sloppy phrase search with the simple query parser by appending the ~ operator and a value to the phrase. For more information, see Searching for Phrases.

Fuzzy search—use the ~ operator to perform fuzzy searches with the simple query parser. Append the ~ operator and a value to a term to indicate how much terms can differ and still be considered a match. For more information, see Searching for Individual Terms.

Highlighting—use the highlight.FIELD parameter to highlight matches in a particular field. For more information, see Highlighting Search Hits in Amazon CloudSearch.

Autocomplete—configure a suggester and submit requests to the suggester resource to get a list of query completions and the documents in which they were found. For more information, see Getting Autocomplete Suggestions in Amazon CloudSearch.

Partial search results—use the partial=true parameter to retrieve partial results when one or more index partitions are unavailable. By default Amazon CloudSearch only returns results if every partition can be queried.

Deep paging—use the cursor parameter to paginate results when you have a large result set. For more information, see Paginate the results.

Match all documents—use the matchall structured query operator to retrieve all of the documents in the index.

New query parsers—use the q.parser parameter to select the Lucene or DisMax parsers instead of the simple or structured parser, q.parser=lucene or q.parser=dismax.

Strings are no longer tokenized on case boundaries and periods that aren't followed by a space are considered part of the term. For more information, see Text Processing in Amazon CloudSearch.

Literal fields are now case-sensitive.

Search responses no longer include the rank, match expression, or CPU time. The only status information returned is the resource ID (rid) and processing time (time-ms).

When you get facet information for an int field, min and max values are no longer returned.

cursor.FIELD

expr.NAME

facet.FIELD

format

fq

highlight.FIELD

partial

pretty

q.options

q.parser

return

sort

boost

matchall

near

phrase

prefix

range

term

bq

facet-FIELD-top-N

facet-FIELD-sort

facet-FIELD-constraints

rank

rank-RANKNAME

return-fields

result-type

t-FIELD

field

filter

-

|

+

*