User Answers

Hi Guillaume,

iFind indices, like bitmap indices before, require a bitmap-friendly ID key (positive integer). When you make a table the child in a parent-child relationship, the underlying storage structure will use a composite key that therefore no longer satisfies the bitmap friendliness. We do plan to lift this limitation in a future release, as it's already the case for bitmap indices, but for now you'll have to review your schema and see if a one-to-many or (preferred) foreign key would work for you.


No, iKnow doesn't extract text from RTF prior to its NLP task. Besides the LibreOffice suggestion, I've also heard people who've worked with simple Java RTF extractors (part of regular JDK) and Tika in the past. 

There is a simple regression calculator that is used internally for similar trend line work, iirc. The class reference is not spectacularly elaborate, but it's fairly straightforward to use. First you use the add() function to load up points and then the result() function will calculate a simple trend line and populate Slope and Intercept properties:

plain variables don't really have a type, so it's worth knowing what you want to use your date for (storing as %Date, %PosixTime, %Timestamp,...). If you're starting from a string and want to convert to IRIS' internal date format, use the $zdateh() function and use 4 as the second argument for European date format. See the docs for that function for more info.

Hi Sean,

IRIS uses different port numbers than Caché and Ensemble so port clashes are not an issue, but there are a few components that are typically shared across instances (e.g. ISCAgent) where consecutive installations of IRIS and Caché might cause trouble. We're documenting these and also other compatibility items of note (such as accessing one platform with the other's xDBC driver) in a guide that will be published shortly.

The SwaggerDoc for the iKnow REST APIs is also available online. Attribute information is relatively fine-grained and associated with entity occurrences aka sentence parts, which you can retrieve per sentence from the /domain/{domain}/sentences/{id}/details endpoint. The /domain/{domain}/entities series of endpoints are for unique entities, with which no occurrence-level information such as attributes is associated.

Hi Robert,

in 2018.2, we're introducing a feature called "coordinated backup", which basically allows adding a checkpoint in the journal files of all participating instances so you can roll them back to a synchronized state. We were just working on the docs for that feature the other week and it's four pages if you'd want the comprehensive answer to your question, so this is just a simplified version :-)