Back, sort of

Wow, it has been a while. Since I posted last, I have picked up a number of new interests and should be able to blog more frequently.

Meanwhile, ...77 hits on March 25 this year?

A link to my current project, Java storage. Nothing spectacular, but not trivial and there are some lessons. Will try to describe what I learned while working on it in a following post.

Video test

Database.com

I have started using database.com. Although the authentication (from Apex Data Loader) is a pain, I have to say I am quite impressed with how powerful the platform is. A couple of quick thoughts, and I will soon have more I hope, as I just started developing an application using it:

• interesting to see how many of the features (value and display lists, row id references) are similar to (for example) Caché 's: basically nothing is ever new in computing it seems
• its usability would be greatly increased I guess if it supported disconnected recordsets so that a client application can use it even without access to the cloud

This also proves that it does make sense in certain instance to maintain one's own platform/language/database stack.

More soon.

Yahoo + Koprol != Love

You would think that it would be easy to link your Yahoo and Koprol accounts.
If only.

Running Cache on a Linux VM

Download SUSE.
• Login: vmplanet
• Pass: vmplanet.net
• RootPass: vmplanet.net

Download the SUSE Cache install from InterSystems.
Install is rather easy with the RPM sw management system on SUSE.
Set the network connection on the VM to "Bridged" to allow the VM to use the host's network.
The ports have to be enabled on the Linux VM:

su root
iptables -F INPUT (Delete all the INPUT rules)
iptables -P INPUT ACCEPT

To enable remote access from the Cache toolset:
• port: 1972
• web port: 57772
• login: ... the usual, _system or SuperUser



To connect from Terminal:
• enable telnet service on VM
• create a .session.sh script to launch Cache session, chmod +x
Cache must already be running (su ccontrol start cache if it is not)
• in .profile for vmplanet user add line to invoke session.sh

More useful appliances.


Studio connected to Cache running on a SUSE VM:




A view from the Linux VM:


Remote terminal connection:

Caché Data Access Modes

Since I have joined InterSystems a while ago, I thought a Caché post might be appropriate, especially since it illustrates quite nicely the various data access modes offered by the platform.
Below is a routine called TestModes.mac, invoked by a >do ^TestModes. The SQLCompile macro is necessary because the SQL code references objects which do not exist compile time, before CreateTable is run (InsertData). The multi-dimensional sparse arrays created as SQL tables (by CreateTable) are accessed in multi-value mode by ReadMV and deleted by DropMV (which also uses the DeleteExtent method of the data objects to perform a ‘table truncation’ on the data objects).
Another nice feature is the extraction of the list items from the arrays (table data rows are stored as array items, whereby each field is stored in a string list created by $LISTBUILD) back to strings by the $LISTTOSTRING function.
Some legacy features are visible (the need to declare SQLCODE as a public variable and make it NEW to restrict it to the current scope). I don’t think that this kind of mix would make sense in an application, where it would make sense to use one access metaphor and only one for a given data item; but this is still a valuable example which ties the data architecture(s) of Caché together nicely.

TestModes
; shows the different data access types
; SQL, MV, class
#SQLCompile Mode = DEFERRED

DO CreateTable
DO InsertData(1000)
DO ReadMV
DO DropMV
WRITE $$ReadData()
QUIT
;
CreateTable()
{
&sql( DROP TABLE items)
&sql( CREATE TABLE items(id INT, data VARCHAR(50)))
QUIT
}

InsertData(upto)[SQLCODE]
{
#DIM x as %String
#DIM i as %Integer

NEW SQLCODE

FOR i = 1:1:upto
{
SET x = i _ " DataItem"
&sql( INSERT INTO SQLUser.items(id, data) VALUES( :i, :x))
IF (SQLCODE '= 0 )
{
WRITE "SQL error:", $ZERROR, !
}
}
QUIT
}

ReadMV()
{
#DIM ky AS %String
#DIM extr AS %String

WRITE "Navigating the data", !

SET ky = $ORDER(^User.itemsD(""))
WHILE( ky '= "")
{
; WRITE ky, ^User.itemsD(ky), !
SET ky = $ORDER(^User.itemsD(ky))
IF ( ky '= "" )
{
SET extr = $LISTTOSTRING(^User.itemsD(ky))
WRITE ?5, extr, !
}
}
QUIT
}

DropMV()
{
DO ##class(User.items).%DeleteExtent()
KILL ^User.itemsD
QUIT
}

ReadData() [SQLCODE]
{
#Dim result AS %String
NEW SQLCODE

&sql(SELECT COUNT(*) INTO :result FROM SQLUser.items )
; the table is still there, but empty

IF (SQLCODE '= 0) SET result = "Error " _ $ZERROR
Q result

}

Tableau Public

My first attempt!

MySQL and SQL Server

...actually, how to make the two talk.

Create a Linked Server in SQL Server (using ODBC):

To query (from SQL Server):

select top 5 * from mysqlanalytics...analytics;

select * from mysqlanalytics...session;

Stored procedures aren't available, but you can get around this by writing a view in MySQL to return the data from the stored procedure (and perhaps use a table to get any parameters the stored procedure might need).

mysql> create view ViewSession as
    -> select GetSession() as CurrentSession;
Query OK, 0 rows affected (0.06 sec)

mysql> select * from viewsession;
+----------------+
 CurrentSession
+----------------+
              1
+----------------+

And ViewSession is now available in SQL Server:

select * from mysqlanalytics...ViewSession