LIKE and IN expressions in parameter queries

  • Thread starter Thread starter Rob Oldfield
  • Start date Start date
John Bailo said:
But that's not what I said.

I'm not trying to replace the dbms with another.

I want to eliminate it entirely and read my data direct as files into c#
and manipulate it there.

WHat I need is a data server simple as an http server that can take raw
requests and let me manipulate on the middle ware.
Click to expand...

Let me get this clear: will you have a server element separate from the
client element or not? If not, then you've eliminated one aspect which
is appealing to many, many people.

If you *are* going to separate them, then either you're suggesting
shunting all the data from one place to another (which would be much
slower in almost all cases than processing the query "in place") or you
*are* effectively replacing the DBMS with another.

It would help if you'd spell out your suggested architecture a bit more
clearly. I would be very surprised if you've got an architecture which
is significantly better than the one(s) the rest of the world uses
though...
 
Jon said:
It would help if you'd spell out your suggested architecture a bit more
clearly. I would be very surprised if you've got an architecture which
is significantly better than the one(s) the rest of the world uses
though...
Click to expand...

My architecture would basically encapsulate very simply file Open/Close
and scan operations, that would message data via http.

There would be no relational aspects.

All the analysis would be done back on a Smart Client...which would take
in needed data as files.

Then it would parse and map out the needed data.

Everything would be done using primitive datatypes -- no object data
until its used for analysis or rendering.
 
John Bailo said:
My architecture would basically encapsulate very simply file Open/Close
and scan operations, that would message data via http.

There would be no relational aspects.

All the analysis would be done back on a Smart Client...which would take
in needed data as files.

Then it would parse and map out the needed data.

Everything would be done using primitive datatypes -- no object data
until its used for analysis or rendering.
Click to expand...

Okay. So in order to process a table which is a gigabyte in size (which
is far from huge), you'd need to transfer the whole gigabyte across the
network, even if only a row or two matches the eventual query?

Um, no, that's not a good way to build a performant system.
 
Back
Top