Custom Query (2765 matches)

Transfer the following two options from rasdl (source:rasdl/rasdl.cc to directql (which is now part of rasserver):

       -c, --createdatabase 
              create database and initialize schema information
       --deldatabase 
              delete database

This will allow to later remove the deprecated rasdl (ticket #1099)

The attached test script​ produces the following output:

N = NULL = 10

-------------------------------------------------------
case 1 - area [0:0, 0:9]:
  slice 1: [N,N,N,N,N,1,1,1,1,1]
  slice 2: [ , , , , , , , , , ],2,2,2,2,2,2,2,2,2,2

max of slice 1 and 2:   10,10,10,10,10,10,10,10,10,10
min of slice 1 and 2:   10,10,10,10,10, 1, 1, 1, 1, 1
  + of slice 1 and 2:   10,10,10,10,10,10,10,10,10,10

-------------------------------------------------------
case 2 - area [0:0, 0:9]:
  slice 1: [N,N,N,N,N,1,1,1,1,1]
  slice 2: [ , , ,3,3,3,3,3, , ],2,2,2,2,2,2,2,2,2,2

max of slice 1 and 2:   0,0,0,10,10,3,3,3,0,0
min of slice 1 and 2:   0,0,0, 3, 3,1,1,1,0,0
  + of slice 1 and 2:   0,0,0,10,10,4,4,4,0,0

-------------------------------------------------------
case 3 - area [0:0, 2:11]:
  slice 1: N,N,[N,N,N,1,1,1,1,1, , ]
  slice 2:  , ,[ ,3,3,3,3,3, , ,2,2],2,2,2,2,2,2,2,2

max of slice 1 and 2:   0,10,10,3,3,3,0,0,0,0
min of slice 1 and 2:   0, 3, 3,1,1,1,0,0,0,0
  + of slice 1 and 2:   0,10,10,4,4,4,0,0,0,0

Both max and min should produce the same results when one of the operands is null, more particularly "if at least one of the operands of a cell operation is null then the overall result of this cell operation is null" (cf. ​documentation). However, in all cases above the null values seem to be treated as normal values, so min and max are different when a null value is present.

It should be replaced with calls to rasserver:

• ​this should become rasserver -c
• ​this should be transformed to several rasserver -q "CREATE TYPE ..." calls, for each type in basictypes.dl.
  • the rasql create type syntax is documented ​here
  • note that this subtask requires to first fix ticket #1859, because it's not possible to create complex cell types with CREATE TYPE currently; the suggested workaround is to implement parser support for CFloat32 (complex) and CFloat64 (complexd) instead in lex.ll / oql.yy