Search
Litium uses Lucene as the underlying search engine. This article explains the general concepts and things to watch out for. Detailed code examples are given in articles under each module.
Litium uses Lucene as the underlying search engine. Developers are not required to know about Lucene when implementing functionality based on search, but some key observations should be kept in mind.
- When a change is made in an entity, it will take a short amount of time for the search engine to index the changes. Therefore, if you search immediately after the change, it might not result in a successful hit.
- Searching has better performance when compared to iterating through a large collection to find an item. For example, if you would like to find all articles with article numbers starting with "001", it is much faster to search.
- When building search, pay attention to the Lucene search text syntax. For example, '*' at the end of a string means a pattern match, with any number of ending characters.
- When searching language is important, because some property values might be language dependent. However, if you are searching for a language independant field, you may use the default language.
- After upgrading or migrating to a new files folder, the search indexes must be rebuilt. This can be done from Litium back office Settings > System Settings > Search Indeces.
In Litium 6 the Lucene search index can be used to search for content in the Sales, Websites and Products areas. In previous versions of Litium (5 and earlier) Lucene can also be used to search for content in the Customers and Media areas.
Avaliable search indexes in Sales:
- ECommerceSearchDomains.Deliveries
- ECommerceSearchDomains.Orders
- ECommerceSearchDomains.Payments
Avaliable search indexes in Websites:
- CmsSearchDomains.Pages
- CmsSearchDomains.PageContents
- CmsSearchDomains.WorkingCopies
Avaliable search indexes in Products:
- ProductCatalogSearchDomains.Categories
- ProductCatalogSearchDomains.Products: The Products index only contains products that:
- have a valid field template
- have a valid display template
- have connected variants.
- ProductCatalogSearchDomains.ProductsAndCategories
This article explains the basic component parts of a search query and shows how to perform a basic search.
It is possible to extract the search results in pages. Paging is important when you do not want to process the entire result set in a single go, and is mostly used when displaying the result in multiple pages to the end user.
This article shows how to obtain search results based on a "Logical OR" operation of two search clauses. It looks more like a two searches done in one go and the results combined.
It is possible to sort the search results in ascending or descending order. This article explains how.
This article describes how to search for a range of values. A typical search in this area would be a price range search or a search for new products published between certain dates.
Tag terms are the values the search index has indexed. It is easy to explain with an example. If we search for all articles that have the value "Leo Tolstoy" for the property field "Author", then the search index "tag name" is "author property field" while its "value term" is "Leo Tolstoy". Note that the "Author" is a property field in the article template, while "Leo Tolstoy" is a value given to this field in a particular article. In other words, for this particular example, "Leo Tolstoy" is a TagTerm for the search tag name "Author property field".
This article explains how to get these tag terms. For example, if you are interested in finding all the "authors" to produce a list of authors of the books you have.
The search filter can be built to exclude certain items by using an exclude tag clause.
If there are multiple hits for a given search string from different property fields, you can influence Lucene to prioritise a particular property field to search for over another, resulting in a prioritised search result. This is called boosting.
For example, if a user searches for the term "Anna" in a book store, that has several books with "Anna" in the title as well as several books written by an author named "Anna". Since the usual case is to search for books by the title, the hits coming from books with "Anna" in the title should be more prominent in the search results, and therefore shown first, and then all other fields should be considered.
Note that boosting may affect the site search performance and therefore should be used only when necessary.
To specify how the search engine handles words/phrases you use analyzers.