update readme's with method used for determing result data type

doc/sql.extensions/README.case

@@ -1,3 +1,5 @@
+SQL Language Extension: CASE
+
 Function:
     Allow the result of a column to be determined by a the results of a 
     case expression.
@@ -6,8 +8,6 @@ Author:
     Arno Brinkman <firebird@abvisie.nl>
 
 Format:
-
-
   <case expression> ::=
       <case abbreviation>
     | <case specification>
@@ -46,9 +46,8 @@ Format:
 
   <result expression> ::= <value expression>
 
-
-Syntax Rules:
-
+Notes:
+    See also README.data_type_results_of_aggregations.txt
 
 Examples:
 

doc/sql.extensions/README.coalesce

@@ -1,3 +1,5 @@
+SQL Language Extension: COALESCE
+
 Function:
     Allow a column value to be calculated by a number of expressions, 
     the first expression returning a non NULL value is returned as the 
@@ -19,6 +21,8 @@ Syntax Rules:
          <case specification>:
          CASE WHEN V1 IS NOT NULL THEN V1 ELSE COALESCE (V2,...,Vn) END
 
+Notes:
+    See also README.data_type_results_of_aggregations.txt
 
 Examples:
 A)

doc/sql.extensions/README.data_type_results_of_aggregations.txt

@@ -0,0 +1,44 @@
+Function:
+    Specify the result data type of the result of an aggregation over values of 
+    compatible data types, such as case expressions and columns at the same 
+    position in a union query expression.
+
+Author:
+    Arno Brinkman <firebird@abvisie.nl>
+
+Syntax Rules:
+
+1) Let DTS be the set of data types over which we must determine the final 
+   result data type.
+       
+2) All of the data types in DTS shall be comparable.
+
+3) Case:
+
+  a) If any of the data types in DTS is character string, then:
+
+    1) If any of the data types in DTS is variable-length character string, 
+       then the result data type is variable-length character string with 
+       maximum length in characters equal to the maximum of the lengths in 
+       characters and maximum lengths in characters of the data types in DTS.
+
+    2) Otherwise, the result data type is fixed-length character string with 
+       length in characters equal to the maximum of the lengths in 
+       characters of the data types in DTS.
+       
+    3) The characterset/collation is used from the first character string
+       data type in DTS.
+
+  b) If all of the data types in DTS are exact numeric, then the result data 
+     type is exact numeric with scale equal to the maximum of the scales of 
+     the data types in DTS and the maximum precision of all data types in DTS. 
+
+  c) If any data type in DTS is approximate numeric, then each data type in DTS 
+     shall be numeric else an error is thrown.
+
+  d) If some data type in DTS is a datetime data type, then every data type in 
+     DTS shall be a datetime data type having the same datetime type.
+
+  e) If any data type in DTS is BLOB, then each data type in DTS shall be 
+     BLOB and all with the same sub-type.
+

doc/sql.extensions/README.derived_tables.txt

@@ -38,9 +38,8 @@ Notes:
     The optimizer can handle a derived table very efficiently, but if the 
     derived table contains a sub-select then no join order can be made (if the 
     derived table is included in an inner join).
-    
 
-Example(s):
+Examples:
 
 a) Simple derived table:
 

doc/sql.extensions/README.nullif

@@ -1,3 +1,5 @@
+SQL Language Extension: NULLIF
+
 Function:
     Return a NULL value for a sub-expression if it has a specific value 
     otherwise return the value of the sub-expression
@@ -10,13 +12,15 @@ Format:
   <case abbreviation> ::=
       NULLIF <left paren> <value expression> <comma> <value expression> <right paren>
 
-
 Syntax Rules:
     1) NULLIF (V1, V2) is equivalent to the following <case specification>:
          CASE WHEN V1 = V2 THEN NULL ELSE V1 END
 
+Notes:
+    See also README.data_type_results_of_aggregations.txt
 
 Examples:
+
 A)
   UPDATE PRODUCTS
     SET STOCK = NULLIF(STOCK,0)