Solr-Suche vs. Indexed Search

Filter? Hit suggestions? Search files? There are different search engines that can be used for the internal search of a website. At wegewerk, we mostly use Indexed Search or Solr Search. We would therefore like to take a closer look at the two variants and get to the bottom of the question: What are the differences between the search engines?
Search functions are extremely helpful for quickly finding what you are looking for on a website. For TYPO3 websites, Solr Search and Indexed Search are often used for this purpose. The TYPO3 Indexed Search is the standard search function of TYPO3 and is supplied as a TYPO3 Core Extension. It only needs to be configured for the respective project. Solr Search, on the other hand, is an external search solution based on Apache Solr. It is significantly more powerful and offers numerous extension options.
For an initial overview, we have compared the main differences between the two search functions in a table:
Feature | TYPO3 Indexed Search | Solr Search |
---|---|---|
Installation |
Simple, integrated in TYPO3, only needs to be configured |
More complex, requires external installation and more extensive configuration |
Performance |
Good for small to medium-sized websites with simple search queries |
Very powerful, ideal for large websites with a lot of traffic and complex search queries |
Scalability | Limited scalability |
High scalability, supports large data volumes and cluster operation |
Search speed |
Basic and sufficient for small amounts of data |
Very fast, optimised for large data volumes |
Relevance ranking* | Basic ranking after indexing |
Advanced ranking, based on algorithms |
Sorting |
Relevance: Sort by keyword occurrence. Date: Newer content at the top. Title: Alphabetical sorting. |
Relevance (score): Extended relevance calculation Date: Sortable by publication date. Otherwise sortable by any field. Multiple sorting: Combined sorting (e.g. relevance + date) possible. |
Faceted search* | Not supported |
Provides filters and faceted navigation |
Extensibility |
Limited, only basic functions available |
High expandability, numerous plugins and customisations possible |
Hit suggestions* | Not supported |
Supports the automatic completion of a search term |
Stop words and synonyms* | Not supported |
Enables the entry of alternative terms and stop words |
Search indexing | Local indexing in the TYPO3 database |
External indexing, very flexible and powerful |
File indexing* |
Supports simple files such as HTML and PDF |
Files can only be indexed via additional extensions (some of which are subject to a charge) |
Multilingualism |
Basically supported, with more restrictions |
Very powerful, supports complex multilingual websites thanks to separation into Solr cores |
Multidomain-capable |
separate indices possible for each domain with limited scalability and fewer configuration options |
Creation of separate indices for different domains or projects possible |
Additional costs (beyond initial configuration) |
none |
Depending on the desired extension and/or additional feature, additional configuration or external costs may arise |
* As the table illustrates, Solr offers more options than Indexed Search. Let's take a closer look at a few of Solr's features below:
What are the special features of Solr?
Faceted search
One of the main arguments in favour of using Solr search is probably the faceted search. A faceted search is the grouping of hits into defined categories. These individual categories are known as facets.
Individual fields or entire field groups can be used as facets. Which fields or field groups are used can vary from project to project. In most cases, one or more of the category groups stored in the system are suitable. We usually make the decision together with our customers during the conception phase. This allows the search results to be narrowed down based on the defined facets. The Solr search also allows several facets to be combined.
Translated with DeepL.com (free version)
Result suggestions
Another function of the Solr search is the simplified input of search terms. Based on the user's input, an attempt is made to anticipate the desired word with the help of automatic completion. These hit suggestions can have different characteristics:
Spelling correction: The ‘did you mean?’ feature suggests alternative search terms for incorrect entries.
Auto-suggestion: Based on the character string already typed, Solr makes search term suggestions as you type.
Similar search terms: Frequently clicked or relevant results are suggested by the top hit feature directly in the search bar.
Stop words and synonyms
In addition to simplifying the search input, Solr also allows the results to be optimised. Synonyms and stop words can be defined or imported in the TYPO3 backend.
Stop words: Stop words have little or no information value in search queries (e.g. and, or, the, that, etc.). These words are removed by Solr during indexing and/or searching in order to better emphasise the relevant terms. Stop words can therefore be an efficient tool to increase the relevance of search results. However, they should be carefully adapted to the requirements of the website and the type of content to ensure that no important terms are inadvertently filtered out.
Synonyms: Synonym lists make it possible to expand search queries by considering alternative terms for certain words. This improves the findability of content, as Solr can also include similar or related terms in the search. Special cases include compound, new or special terms that are not in the dictionary or alternative spellings of words. Examples of such terms are: wegewerk, Wegewerk, ww or TYPO3, Typo3, typo3, t3, T3.
Why is this important?
If the synonyms are not specified, each variant is technically a separate ‘search result’. The system treats the terms as two different words, even if they logically mean the same thing. As a result, a search for ‘wegewerk’ will return different results than a search for ‘Wegewerk’.
To avoid this, synonyms should be created editorially for these special words so that Solr summarises the search results correctly.
File contents
Simple files such as PDF and HTML can be indexed in the Indexed Search. To be able to search files and their metadata in Solr, however, a separate extension for TYPO3 is required. This allows linked files or entire folders to be indexed in Fileadmin. There are various extensions (free and paid) to choose from. We will be happy to advise you on which extension makes the most sense for your project.
Relevance criteria and boosting
The ranking of the search results also differs between the two variants. While Indexed Search carries out a basic ranking after indexing, the relevance of search results in Solr is primarily calculated using ranking algorithms. These determine the significance or importance of a page in relation to a search query. How often a term appears on a page and how rare a term is in the overall index plays a role here.
Diese Relevanzbewertung von Solr kann durch das sogenannte Boosting verändert werden. Das Boosting von Suchergebnissen erhöht die Relevanz bestimmter Dokumente oder Felder in der Suchergebnisliste, basierend auf vordefinierten Kriterien. Dies kann gezielt genutzt werden, um bestimmte Inhalte in den Suchergebnissen höher zu gewichten, indem ihnen ein höherer Boost-Wert zugewiesen wird.
There is the option of boosting at field level, document-based boosting or function-based boosting (e.g. popularity, ratings or date).
No explicit boosting is used in the standard configuration of Solr. All fields and documents are therefore initially weighted equally. To activate boosting, this must be done explicitly in the configuration or in the search query. If this is desired, it makes sense to define specific use cases in advance, which can be used to test and refine boosting.
Solr and Indexed Search - Which search is suitable for which purpose?
Depending on the complexity of your own website and the existing requirements for the search function, the use of Solr or Indexed Search is more suitable. So which search should you use?
Solr: Solr is suitable if a large number of search results need to be filtered. Usually, facets such as categories, type of result or time periods are used. Extensive pages with lots of (different) content in particular benefit from the powerful search engine.
Indexed search: Indexed search, on the other hand, is suitable for small to medium-sized websites with a manageable number of search results. The relevance search functions, which are preset in Indexed Search, are often sufficient here.
We will be happy to advise you on which search engine is best suited to your project.
Tips and tricks for using Solr - answers to the most frequently asked questions about the Solr search:
Why is my PDF not being found?
To find files in TYPO3 using Solr, an additional extension is required. This extension must first be installed and set up.
The PDF must also be publicly accessible.
Why is my page/content not found?
Especially with newly created pages or content, it can happen that the server has not yet had time to index, i.e. search, the content.
How long it takes to index the page depends on various factors and cannot be determined exactly. For large pages, the waiting time can be an hour or longer.
Why is the expected search result not at the top of the list?
Solr uses a ranking algorithm. If the search results do not match the expected results, these ranking factors can be customised with boosting settings.