gudusoft.gsqlparser.pp2.token

Class DoubleQuotedStrings

Object

gudusoft.gsqlparser.pp2.token.DoubleQuotedStrings

public final class DoubleQuotedStrings
extends Object

Which double-quoted texts may be character strings rather than delimited names. Changing the case of a delimited name is a formatting choice; changing the case of a string changes the data. So the formatter leaves every double-quoted text this class reports as written, and pp2's content checks refuse any change to one.

A double-quoted text may be a string

• anywhere, in a dialect whose identifier rules do not make a double quote a name delimiter (MDX, for one), or in which a setting the SQL does not show can make it a string: MySQL and its relatives without ANSI_QUOTES, BigQuery, Hive, Spark, Databricks, Impala, Flink, Couchbase, Informix without DELIMIDENT, Sybase with quoted_identifier off (its default), SQLite (which falls back to a string), SOQL, and Access (whose criteria take "Smith");
• after the statement that turns the quoted-identifier option off: SET QUOTED_IDENTIFIER OFF, SET ANSI_NULLS, QUOTED_IDENTIFIER OFF or SET ANSI_DEFAULTS OFF (which sets QUOTED_IDENTIFIER with it) in SQL Server, Azure SQL and SQL Anywhere, SET [TEMPORARY] OPTION quoted_identifier = 'Off' in SQL Anywhere, and SET SQL DIALECT 1 in Firebird, whose dialects before 3 read double quotes as strings.

Only the SQL text is seen. A setting it does not show is the dialect's default, which is how GSP's parser reads the text too: in SQL Server and Azure SQL QUOTED_IDENTIFIER is on (Microsoft's ODBC and OLE DB drivers turn it on when they connect), in SQL Anywhere quoted_identifier is on, and Firebird uses dialect 3. A script for a connection that has the option off must turn it off itself; without that, its double-quoted texts are names here.

A later statement that turns the option back on is not followed. In a SQL Server routine body it does not change how the body is read (the setting of the batch that created the routine does), and the tokens alone cannot tell a routine body from a batch; leaving a name's case as written is safe, re-casing a string is not. A script that never turns the option off keeps the identifier reading.

Field Summary

Fields

Modifier and Type	Field and Description
static long	ALL Every double-quoted text may be a string.
static long	NONE No double-quoted text may be a string.

Method Summary

Modifier and Type	Method and Description
static long	stringsAfter(EDbVendor vendor, Pp2TokenStream stream) stringsAfter(EDbVendor, TSourceTokenList) over a pp2 token stream.
static long	stringsAfter(EDbVendor vendor, TSourceTokenList tokens) The offset after which every double-quoted text of tokens may be a string: ALL, the offset of the value that turns the quoted-identifier option off, or NONE.

Methods inherited from class Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

NONE

public static final long NONE

No double-quoted text may be a string.

See Also:

Constant Field Values

ALL

public static final long ALL

Every double-quoted text may be a string.

See Also:

Constant Field Values

Method Detail

stringsAfter

public static long stringsAfter(EDbVendor vendor,
                                TSourceTokenList tokens)

The offset after which every double-quoted text of tokens may be a string: ALL, the offset of the value that turns the quoted-identifier option off, or NONE. A double-quoted token may be a string when its offset is greater.

stringsAfter

public static long stringsAfter(EDbVendor vendor,
                                Pp2TokenStream stream)

stringsAfter(EDbVendor, TSourceTokenList) over a pp2 token stream.