How to import user-defined Data from Database or TCP/IP server

[TinTin82]

Hi,
I’m very new in root programming and i want use root for visualization of forex market data.
I have a big Database with a self-made structure, which holds the forex data for example EUR/USD.

Is it possible to read in data from a odbc database which have a user-defined structure ??

Can i read in data into root from a TCP/IP server ??

[linev]

Hi,

There are ODBC-support classes in ROOT

https://root.cern/doc/master/classTODBCServer.html#a4987eb3b668c79b453df5c847cd22f68

You can request any kind of information, using standard SQL syntax.
Here example macro how SQL classes (including ODBC) can be used:

root-project/root/blob/master/tutorials/sql/sqlselect.C

There is also TSocket class, while allows to open server and client sockets.

https://root.cern/doc/master/classTSocket.html

Examplse you will find in tutorials/net like hserv.C or hclient.C

Regards,
Sergey

[TinTin82]

Hi Linev,

cool, now i can connect to my Server. But i have written my server with windows sockets and i transmit a data structure.

I have imported the structure in root and noticed that the structure alignment is different from the alignment in my server.

Visual Studio 2013 Server:

typedef struct { … }s_TCP_IP_SEND_BUFF_t;
…
iResult = send(ClientSocket[ui32_Counter], (char *)&s_TCP_IP_SEND_BUFF, sizeof(s_TCP_IP_SEND_BUFF), 0);

Root 5.34/36 Client

typedef struct { … }s_TCP_IP_SEND_BUFF_t;
…
int ret = sock->RecvRaw((void *)(&s_TCP_IP_SEND_RECV), sizeof(s_TCP_IP_SEND_RECV_t));

[linev]

@TinTin82

I have imported the structure in root and noticed that the structure alignment is different from the alignment in my server.

Simplest solution for alignment - use data types which are 8-bytes aligned. Like int64_t or double, but not bool.

[TinTin82]

This type in one system is 16 bytes, and in the other it is 12 byte.
>typedef struct s_test_t
{
uint8 ui8_v1;
uint64 ui64_v2;
}s_test_t;

I started to add padding bytes like
>typedef struct s_test_t
{
uint8 ui8_v1;
uint32 ui32_PADDING_0;
uint64 ui64_v2;
}s_test_t;

to make it the same size. The reason is that i must adapt the root types to the server types. It’s not possible to
adapt the server in this case.

[linev]

@TinTin82

The reason is that i must adapt the root types to the server types. It’s not possible to

adapt the server in this case.

If you cannot change server code, you can use #pragma pack as described here. Code will look like:

#pragma pack(push, 1) // exact fit - no padding
struct MyStruct
{
   char b; 
   int a; 
   int array[2];
 };   
 #pragma pack(pop) //back to whatever the previous packing mode was

But most probably, you will need to run macro in compiled mode.