Open subtreeSearch Close subtreeAdvanced Search   
     Resources
     Search Terms
     Content Type
     Document Date
     Geography
     Sorting
     Exact Phrases

Open subtreeDownloads Accessibility

Exact Phrases

What exactly is an excact phrase search?

The Advanced Search form provides radio buttons for modifying the way in which the submitted search terms are to be evaluated. One of the radio buttons is reserved for the Exact phrase search.

The exact phrase is different from the other kinds of searches. The search terms are not broken down into words. Rather, the whole expression, as entered into the text field, is used for exact phrase matches in the target pages. This includes any special characters such as commas, quotes, hyphens, white spaces etc. The underlying MySQL database looks for precise substring matches and only returns pages which have one or more such hits.

Important note

An exact phrase is normally much slower than a normal search because the underlying MySQL database cannot use any fulltext indexes for this kind of a search. Also, the quality of the OCR for many pages is fuzzy, hence an exact phrase search may not always return the expected results.

Differences between exact phrase searches and quoted phrases

The exaxt phrase search, with the "Exact phrase" radio button selected, uses the whole expression, as it was originally entered into the text field. It does not provide any special treatment for phrases with surrounding quotes. Rather, if the text field includes quotes, these are then included as part of the exact phrase, just as it includes any other special characters.

By contrast, the other search kinds where the radio button selected is one of the "All of the words", "Any of the words", or "Boolean expression", treat quoted phrases differently. First of all, the surrounding quotes are not part of the phrase itself. Secondly, all special characters and white spaces (that is all characters other than letters, digits, single quotes and underscores) are removed before executing the search. It then breaks down the remaining contents into the words and attempts to find any instances where the exact sequence of words appears in the target pages or the metadata. If the words from a target sequence are separated by special characters and whitespaces, then these are ignored the same way as in the submitted search terms. The underlying MySQL database uses fulltext indexes for finding the word sequences. Therefore, this kind of a search is much faster than the exact phrase search and often delivers better results.