On 3/19/2017 11:29 AM, donato wrote: > Well, if I change it to text, then I get 0 results...
Did you completely reindex (deleting the index beforehand) after you changed from string to text? If you didn't, then nothing's going to work. Most schema changes require a reindex. https://wiki.apache.org/solr/HowToReindex > I seriously am at a loss... I don't know why this is so hard. > Everything else works, except for this! Something that should be SOOO > simple. Why am I not getting this!? Everything I have tried does not > work. It just doesn't! One thing you were told earlier is that each vertical section of the analysis screen shows the INPUT to that filter. This is incorrect. Each section shows the OUTPUT from the filter that is described by that section. The input is shown BEFORE each section. I am absolutely certain about this. Building a schema that does what you want can be a slow and sometimes painful process, until you learn what filters are available and how they interact with each other. The analysis screen can allow you to experiment with an alternate fieldType, one that doesn't affect your running index, but verifying that it is working 100 percent correctly with live data will involve changing the schema entries for one or more existing fields and performing a reindex. Removing the possessive 's from words requires making a fieldType that has an aggressive tokenizer which is capable of producing separate terms for each word. If the tokenizer doesn't support removing english possessives (I do not think any of them do this), then you will need a filter to accomplish that goal. The english possessive filter has already been mentioned in this thread, but seems to be missing from the reference guide. The WordDelimiterFilter can also perform this task, along with a lot of other functionality that may or may not be useful. https://cwiki.apache.org/confluence/display/solr/Filter+Descriptions#FilterDescriptions-WordDelimiterFilter There are others: https://cwiki.apache.org/confluence/display/solr/Filter+Descriptions#FilterDescriptions-ClassicFilter General Note: It is recommended that you do not make large changes to the fieldType entries that you find in an example schema. Instead, create NEW fieldType entries when you need something very different, so that people who are trying to help will not be confused when a type name they're familiar with doesn't behave the way they expect. For example: The "string" type is well-known to be a type that only permits *EXACT* full-input matches or wildcard matches, and cannot have an analyzer. If you modify the string type to solr.TextField and add an analyzer, it will behave very differently than people on this list will expect. Thanks, Shawn
