Packing intervals

Solution to the "packing intervals" problem by Itzik Ben-Gan:
http://www.solidq.com/sqj/Pages/2011-March-Issue/Packing-Intervals.aspx

Scroll to code listing 4 for the SQL code.

Registration is required to copy and paste the code. It can only be copy and pasted from the PDF that can be downloaded after registration.

Execute query under cursor

There is no way in SQL Server Management Studio (SSMS) to execute the query under the cursor. There is the option to execute the entire script with F5 or CTRL+E, but the script might contain multiple SQL statements.

The quickest solution is to install a freeware add-in:

• Download dbForge SQL Complete:
  Note: registration is required
• Install the add-in

The option for auto-capitalization and intellisense is enabled by default. This can be disabled, so only the "execute query under the cursor" (CTRL+SHIFT+E) feature remains:

• Start SMSS
• SQL Complete -> Options -> General -> uncheck "Enable SQL Complete"

Count of weekend days in date range

Use the formula in "calc_name" or "calc_df" to calculate the number of inclusive weekend dates in a date range. Inclusive means that if e.g. the start date is a Saturday, it is counted as 1. The code is pure SQL Server T-SQL, with no need for helper tables or cursors.

Note:

• "Set datefirst" is not needed, but can be used to check the correct outcome of the "calc_df" calculation
• The "calc_df" function is more robust, since it does not rely on localized day names. But it is a bit more verbose and obfuse then "calc_name"
• Column "correct" contains the correct number of weekend dates that can be used as a reference for the calculation
• Subquery X contains some random test values for start- and enddates

SET DATEFIRST 7

SELECT
x.*,
(DATEDIFF(wk, sd, ed) * 2)
  +(CASE WHEN DATENAME(dw, sd) = 'Sunday' THEN 1 ELSE 0 END)
  +(CASE WHEN DATENAME(dw, ed) = 'Saturday' THEN 1 ELSE 0 END) as calc_name,
DATEDIFF(wk, sd, ed) * 2
  +CASE DATEPART(dw, sd)+@@datefirst WHEN 8 THEN 1 ELSE 0 END
  +CASE DATEPART(dw, ed)+@@datefirst WHEN 7 THEN 1 WHEN 14 THEN 1 ELSE 0 END as calc_df,
@@DATEFIRST as datefirst,
(DATEDIFF(wk, sd, ed) * 2) as weeks,
CASE WHEN DATENAME(dw, sd) = 'Sunday' THEN 1 ELSE 0 END as sun_start,
CASE DATEPART(dw, sd)+@@datefirst WHEN 8 THEN 1 ELSE 0 END as sun_start_df,
CASE WHEN DATENAME(dw, ed) = 'Saturday' THEN 1 ELSE 0 END as sat_end,
CASE DATEPART(dw, ed)+@@datefirst WHEN 7 THEN 1 WHEN 14 THEN 1 ELSE 0 END as sat_end_df,
DATEPART(dw, sd)+@@datefirst as sun_df,
DATEPART(dw, ed)+@@datefirst as sat_df
from
(
select 1 as correct, '2013/3/17' as sd, '2013/3/22' as ed union
select 2 as correct, '2013/3/16' as sd, '2013/3/22' as ed union
select 2 as correct, '2013/3/16' as sd, '2013/3/17' as ed union
select 3 as correct, '2013/3/16' as sd, '2013/3/23' as ed union
select 3 as correct, '2013/3/10' as sd, '2013/3/22' as ed
) x

Quipu JDBC connection to MS Access

There are several ways to connect to a MS Access database from Java, like the low-level MS Access library called Jackcess: http://jackcess.sourceforge.net/

The downside is that most Java applications that support heterogeneous connections have a default built-in procedure to connect and open the database. Most of the time, the "getSchemanames" function is called and this one doesn't function as expected with MS Access. Quipu uses this method also.

Therefore, a connection can be made in the following way, but results in an error when retrieving the schema names during reverse engineering with Quipu from MS Access

1) JDBC to ODBC bridge

Create a connection to the MS Access database d:\test.mdb using a native and DSN-less JDBC to ODBC bridge:

Manage connection types:
Add a new connection type:

name= ms access jdbc
type= jdbc
database or file delimiter=
database or file required= false
jdbc port delimiter= :
jdbc default port=
jdbc driver= sun.jdbc.odbc.JdbcOdbcDriver
url type= jdbc:odbc:Driver={Microsoft Access Driver (*.mdb, *.accdb)};DBQ=

Manage connections:
Add a new connection:

name: ms access jdbc
database= test.mdb
hostname= d
port= /

Notes:

• The url type above is for Java x64 and MS Access x64. If Java x32 is used, then the url type= jdbc:odbc:Driver={Microsoft Access Driver (*.mdb)};DBQ=
• No additional jdbc driver (jar file) needs to be installed, because the driver is embedded in the Java JRE runtime

There is an alternative method:

2) UCanAccess

It is a open source database driver written by Marco Amadei that behaves identical to the JDBC to ODBC bridge, but
- doesn't use ODBC
- handles the getschemanames call, using a default "PUBLIC" schema

Download it here: http://sourceforge.net/projects/ucanaccess

To use it with Quipu:
- Extract the jar files from the \lib folder in the zip file to Quipu's \webapps\qp\WEB-INF\lib folder
- Sort files by name and make sure no duplicates exist, e.g. hsqldb and hsqldb-1.8.0 is not allowed.
- Remove the duplicate with the lowest versionnumber. In the case of hsqldb, which has no versionnumber: keep the one from the UCanAccess zipfile

Manage connection types:

Add a new connection type:
name= ucanaccess
type= jdbc
database or file delimiter=
database or file required= false
jdbc port delimiter= :
jdbc default port=
jdbc driver= net.ucanaccess.jdbc.UcanaccessDriver
url type= jdbc:ucanaccess://

Manage connections:

Add a new connection:
name: ucanaccess
database= test.mdb;showschema=true
hostname= d
port= /

Notes:

- The "showschema=true" property bypasses the getSchemanames error