Abstract: Multiple search indexes can be served from a common set of resources. Instead of requiring a processor to be dedicated to serving a single search index, a processor can provide responsive documents for search queries that are based on different ranking algorithms and/or different sets of documents.

Abstract: A unified fresh search result serving approach that employs instant cache update. The instant result cache update efficiently updates the cache server instantly with the latest version of results in time. A unified cache service stores all related documents; thus, all information of a web document is saved on the same server to easily merge all the related information of web documents when generating captions for that document. Any update to web document content is immediately available in the result.

Abstract: Multiple search indexes can be served from a common set of resources. Instead of requiring a processor to be dedicated to serving a single search index, a processor can provide responsive documents for search queries that are based on different ranking algorithms and/or different sets of documents.

Abstract: Optimized storage and query of multi-value properties is enabled by generating a base table for single value properties associated with a user-defined schema and a number of row-based tables for each different type of multi-value properties. By combining multi-value properties into row-based tables or even into compatible type tables, a number of tables to be created for a database system such as SQL system is reduced at the same time optimizing create/read/update/delete operations.

Abstract: A classifier is used to predict relevant results with arbitrary filtering conditions specified by the user. The classifier model is stored as a database table and joined with a metadata properties table instead of calculating the query result probability using the full classifier model. A user-specified query based filter is applied to the joined tables to obtain the list of documents satisfying the filter. The probability is then computed using the sub-model.

Abstract: Methods, systems, and media are provided for synchronizing information across multiple environments of a synchronization system. A search query is received into a frontend infrastructure of a first synchronization environment. The frontend infrastructure checks a local cache manager to see if results already exist for the search query. If existing results are not found, then one or more backend search engines of the first synchronization environment are utilized for the search query. The search results from the backend search engines are saved into the local cache manager of the first synchronization environment. A cache sync notification is created to identify the contents and location of the actual saved results. The cache sync notification is saved in a cache synchronization service located within the first synchronization environment, and broadcast to all other synchronization environments within the synchronization system. The actual results can be retrieved from any other synchronization environment.

Abstract: To provide data from a database in response to an application query, the database is queried an initial time to provide a first set of data. For example, a “fast first page” may be requested, by requesting that all data which might appear on a first page to be displayed is obtained from the database first. After this, one or more subsequent asynchronous queries obtain the remaining information responsive to the query. Or, certain columns from the database (“spine columns”) may be requested in the initial query of the database, and the rest of the data requested subsequently. Thus, information needed to make an initial presentation to a user can be obtained first, and other information filled in subsequently.

Abstract: Optimized storage and query of multi-value properties is enabled by generating a base table for single value properties associated with a user-defined schema and a number of row-based tables for each different type of multi-value properties. By combining multi-value properties into row-based tables or even into compatible type tables, a number of tables to be created for a database system such as SQL system is reduced at the same time optimizing create/read/update/delete operations.

Abstract: A classifier is used to predict relevant results with arbitrary filtering conditions specified by the user. The classifier model is stored as a database table and joined with a metadata properties table instead of calculating the query result probability using the full classifier model. A user-specified query based filter is applied to the joined tables to obtain the list of documents satisfying the filter. The probability is then computed using the sub-model.

Abstract: To provide data from a database in response to an application query, the database is queried an initial time to provide a first set of data. For example, a “fast first page” may be requested, by requesting that all data which might appear on a first page to be displayed is obtained from the database first. After this, one or more subsequent asynchronous queries obtain the remaining information responsive to the query. Or, certain columns from the database (“spine columns”) may be requested in the initial query of the database, and the rest of the data requested subsequently. Thus, information needed to make an initial presentation to a user can be obtained first, and other information filled in subsequently.