Valid Names for Datasets, Columns, Aliases, Paths and Functions

Both RAQL and EMML have the following rules for valid column or alias names and the names within paths to dataset rows:

Column, alias, path and function names may contain letters, numbers and the underscore (_) character. They cannot contain:

Common punctuation or special characters such as $, . (period), - (dash) or : (colon).

There are two exceptions to rule for using periods in names:

Function names use a period to separate the library name for the function from the function name. For example:

Dataset names can be prepended to column names with a period to clarify the exact context for a column in queries where multiple datasets are used. For example:

Column, alias or function names must begin with a letter. A column name of 2000 is not a valid name, but Y2000 is.

The names in paths used with XML datasets must follow XML name rules. They can contain letters, numbers and underscores(_).

Column and alias names are case insensitive and must be unique within a dataset. For example, a dataset row could not contain both a column named item and another named Item.

Column, alias, function and path names cannot be the same as any RAQL Reserved Keywords.

Functions for RAQL queries are written in Java and thus must also follow Java name requirements.

There are work arounds to handle some invalid name issues. See Fixing Invalid Names for instructions.

Names for datasets are the names of the EMML variables that hold the dataset stream within a mashup. Dataset names have the same character and reserved keyword restrictions as column, alias and path names (See Column, Path, Alias and Function Names and RAQL Reserved Keywords).

They must also be unique within the mashup, but dataset names are case sensitive. Although it is not a good practice, a mashup could have a dataset named events and another named EVENTS.

RAQL automatically fixes some issues in invalid names for datasets. With CSV datasets, for example, RAQL will replace spaces in column names with an underscore (_) or change numeric column names to column_original-number.

You can also enclose invalid names in double quote marks ( " ) to fix many invalid name problems.

Note:

You cannot use single quote marks (') to enclose invalid names because this indicates the string is a literal value rather than a name. See Literal Values in Conditions or Expressions for more information.

For example, this query will run even though three of the columns names are invalid or match reserved keywords:

select "dash-column", "dot.column", "date", valid_column from myDataset