Update Patterns Decoupling SQL.md

pchemguy · pchemguy · commit 4c1b13155245 · 2022-07-25T22:11:30.000+03:00
diff --git a/Patterns Decoupling SQL.md b/Patterns Decoupling SQL.md
@@ -14,4 +14,4 @@ SQLite does not support stored procedures, and the application is responsible fo
 
 A better strategy is to combine SQL snippets into a source code library represented by textual modules containing SQL code only. The application should load this library at startup and parse it (SQL comments in the source library can hold the necessary metadata along with inline documentation). Alternatively, the application may store SQL snippets in a database table, such as *queries(name, sql, doc, ...)*. A dedicated application routine would query this table, returning an SQL snippet based on its name. This routine is the only potential exception for inlining basic SQL, responsible for accessing the *queries* table.
 
-With either storage approach, such an SQL library puts all SQL code in compact SQL-only form, facilitating the development process. It is essential, however, that both the calling application and SQL rely on parameterized queries rather than dynamic SQL generation by the application. Parameterized queries, like stored procedures, can present black boxes to the application with a similar SQL interface, which involves the query label, query parameters, and returned columns. As long as the SQL interfaces are stable, the application code can use a conventional query label/name to identify the appropriate snippet (in either file or database library) and send it verbatim to the database. Such an SQL library may be reused, for example, on different platforms or, where relevant, with multiple API bindings.
+With either storage approach, such an SQL library puts all SQL code in compact SQL-only form, facilitating the development process. It is essential, however, that both the calling application and SQL rely on parameterized queries rather than dynamic SQL generation by the application. Parameterized queries, like stored procedures, can present black boxes to the application with a similar SQL interface, which involves the query label, query parameters, and returned columns. As long as these SQL interfaces are stable, the application code can use a conventional query label/name to identify the appropriate snippet (in either file or database library) and send it verbatim to the database. Such an SQL library may be reused, for example, on different platforms or, where relevant, with multiple API bindings.

Original file line number	Diff line number	Diff line change
`@@ -14,4 +14,4 @@ SQLite does not support stored procedures, and the application is responsible fo`
`14`	`14`
`15`	`15`	A better strategy is to combine SQL snippets into a source code library represented by textual modules containing SQL code only. The application should load this library at startup and parse it (SQL comments in the source library can hold the necessary metadata along with inline documentation). Alternatively, the application may store SQL snippets in a database table, such as queries(name, sql, doc, ...). A dedicated application routine would query this table, returning an SQL snippet based on its name. This routine is the only potential exception for inlining basic SQL, responsible for accessing the queries table.
`16`	`16`
`17`		-With either storage approach, such an SQL library puts all SQL code in compact SQL-only form, facilitating the development process. It is essential, however, that both the calling application and SQL rely on parameterized queries rather than dynamic SQL generation by the application. Parameterized queries, like stored procedures, can present black boxes to the application with a similar SQL interface, which involves the query label, query parameters, and returned columns. As long as the SQL interfaces are stable, the application code can use a conventional query label/name to identify the appropriate snippet (in either file or database library) and send it verbatim to the database. Such an SQL library may be reused, for example, on different platforms or, where relevant, with multiple API bindings.
	`17`	+With either storage approach, such an SQL library puts all SQL code in compact SQL-only form, facilitating the development process. It is essential, however, that both the calling application and SQL rely on parameterized queries rather than dynamic SQL generation by the application. Parameterized queries, like stored procedures, can present black boxes to the application with a similar SQL interface, which involves the query label, query parameters, and returned columns. As long as these SQL interfaces are stable, the application code can use a conventional query label/name to identify the appropriate snippet (in either file or database library) and send it verbatim to the database. Such an SQL library may be reused, for example, on different platforms or, where relevant, with multiple API bindings.