This project is read-only.

1. Data Types

HubbleDotNet supports following data types:

Int, TinyInt, SmallInt, BigInt, Float, DateTime, SmallDateTime, Date, NVarchar, Varchar, NChar, Char

DateTime is same as DateTime in .net, measured by milliseconds

SmallDatetime ranges from 1980.01.01 to 2047.12.31 23:59:59, measured by seconds

Date is from 0001.1.1 to 5883516.1.1, measured by days

nvarchar, varchar, nchar, char are same as the four types in SQL SERVER 2005. User can specify length or use MAX

if using above types as Untokenized index, maximum length is 32 characters.

Int, TinyInt, SmallInt, BigInt, Float are same as that are in SQL SERVER 2005. These types can only be used as None or Untokenized index. They can not be used as tokenized index.

column name definition is similar to that in TSQL. The only difference is that we use single quote to escape keywords eg: 'where'. In TSQL it would be [where].


2. HubbleDotNet Index

HubbleDotNet provides three types of indices

Tokenized Index

Tokenized Index is full text index. Only character type can be used as Tokenized Index. Index engine will do work segementation to the text in those columns, then build Inverted Index on the result. Then searching, it will search Inverted Index to find corresponding record.

Untokenized Index

Untokenized Index is a key-value index. All data types can be used as key-value index. Index engine will build index according the key values, and save them in Payload.db file. This index will only be used when search full text and meta data, eg doing time related sorting while searching full text.

None Index

If a column is set to None. HubbleDotnet won't create index for this column. If search condition contains such column, HubbleDotNet will use the index in relational database for query. User needs to create index in relational database for such columns.

Last edited Aug 8, 2011 at 7:54 AM by linkspeed, version 1


No comments yet.