Knowledgebase

Searching Text in Databases Print

  • 0

Finding words in content.

WHAT PATTERN MATCHING PROVIDES

Simple matching, usable with an index only without a leading wildcard.

WHY A LEADING WILDCARD IS FATAL

It forces a scan of every row.

WHAT FULL-TEXT INDEXING PROVIDES

Word-based searching, with relevance ranking.

WHAT IT HANDLES

Multiple words Stemming, in some configurations Ranking by relevance

WHAT IT DOES NOT HANDLE WELL

Misspellings Synonyms Very short words, frequently excluded Languages it was not configured for

WHAT MINIMUM WORD LENGTH SETTINGS CAUSE

Short terms returning nothing, confusingly.

WHAT STOP WORDS ARE

Common words excluded from the index.

WHY THAT SURPRISES USERS

Searching for a common phrase returns nothing useful.

WHAT A DEDICATED SEARCH ENGINE PROVIDES

Far better relevance, tolerance of misspelling, faceting and scale.

WHAT IT COSTS

Another system to operate, and keeping it synchronised.

WHEN TO MOVE TO ONE

When search quality is a product feature rather than a convenience.

WHAT TO DO BEFORE THEN

Use full-text indexing properly, and set expectations.

WHAT TO ALWAYS ENFORCE

That search results respect permissions.

WHY

Search is a common route to exposing data the interface hides.

WHAT TO MONITOR

Searches returning nothing.


Was this answer helpful?
Back

Are you happy with your experience? Leave us a review on Trustpilot.


Trustpilot