monkeymike.co.uk

Install Eclipse under ubuntu 12.04

admin — Sun, 20 May 2012 12:52:52 +0000

Install (just extract the gz file to your home drive if you want) the version on the Eclipse web site, then create a file called eclipse.desktop and give it the following contents:

#!/usr/bin/env xdg-open
 
[Desktop Entry]
Version=1.0
Type=Application
Terminal=false
StartupNotify=true
Icon=/home/mike/eclipse/icon.xpm
Name=Eclipse Indigo 3.7
#Exec=env UBUNTU_MENUPROXY=0 /eclipse
Exec=env /home/mike/eclipse/eclipse

making sure you put the correct path in the two places you need it (the execution path and the image path)

New Fedora coming soon

admin — Thu, 17 May 2012 22:16:59 +0000

…with possibly the un-coolest name ever.

Are you looking at my pint?

mike — Fri, 11 May 2012 20:32:46 +0000

No wonder people felt the Germans were a little aggressive in the late 30s and early to mid 1940s

Oh deer

mike — Fri, 11 May 2012 06:44:05 +0000

Its taken them all week to notice that we were here but they have finally arrived. Welcome the fat deer!

Eat my seed, fat deer!

Assorted pics

mike — Tue, 08 May 2012 18:45:58 +0000

Holiday leaf

And random pattern from the council

Small fossil

mike — Tue, 08 May 2012 16:27:58 +0000

Pulled this tiny one out of the mud and brought it home

A long way to come

mike — Sun, 06 May 2012 16:02:16 +0000

But it’s like:£1.429 per litre of diesel
Thats a saving of £1.20 on a full tank for the cost of ~£43

Arrived

mike — Sat, 05 May 2012 15:08:00 +0000

After a 5 or 6 hour journey without traffic problems we have blebbed out the excessive contents of the car into the chalet. And i am mighty glad to do it. The family were completely well behaved.

Calculate Percentiles with MS-SQL 2005 using interpolation…

admin — Thu, 03 May 2012 06:53:20 +0000

You’ll need MS SQL 2005 or above, it looks like version 2012 has it built in see here!
This is my first use of the OVER bit so it may be a little ugly, also I have been using the old way to pivot stuff so it’s a bit messy at the end.

declare @t table (percentile numeric(10,5))
insert into @t values (0)
insert into @t values (0.25)
insert into @t values (0.5)
insert into @t values (0.75)
insert into @t values (0.8)
insert into @t values (1.0)
 
--give rownumbers for your data over a partition
drop table ##mgw
SELECT 
	s.gppraccode
	,age
	,ROW_NUMBER() OVER (partition by gppraccode ORDER BY age) AS "RowNumber"
INTO ##mgw
FROM mgw_ane_base AS s 
 
--calculate the individual values per percentile
drop table ##mgw_percentileHolding
select	t.percentile*(max(rownumber)-1)+1 as act
		,floor(t.percentile*(max(rownumber)-1)+1) as K
		,((t.percentile*(max(rownumber)-1)+1) % 1) as D
		,count(*) as N
		,t.percentile
		,a.gppraccode
into ##mgw_percentileHolding
from ##mgw as a join @t as t on 1=1
group by t.percentile,a.gppraccode
 
--fold it all over in a manual-ish pivot
select 
	gppraccode
	,max(case when percentile=0 then PercentileValue end) as [0%]
	,max(case when percentile=0.25 then PercentileValue end) as [25%]
	,max(case when percentile=0.5 then PercentileValue end) as [50%]
	,max(case when percentile=0.75 then PercentileValue end) as [75%]
	,max(case when percentile=0.8 then PercentileValue end) as [80%]
	,max(case when percentile=1 then PercentileValue end) as [100%]
	,N
FROM
	(
	SELECT 
	*
	,(SELECT max(age) from ##mgw as b where b.rownumber<=a.act and a.gppraccode=b.gppraccode) as below
	,(SELECT min(age) from ##mgw as b where b.rownumber>=a.act  and a.gppraccode=b.gppraccode) as above
	,(SELECT min(age) from ##mgw as b where b.rownumber>=a.act  and a.gppraccode=b.gppraccode)-(select max(age) from ##mgw as b where b.rownumber<=a.act and a.gppraccode=b.gppraccode ) as f
	,case	when percentile=0 then (SELECT min(age) from ##mgw as b where a.gppraccode=b.gppraccode) 
			when percentile=1 then (SELECT max(age) from ##mgw as b where a.gppraccode=b.gppraccode)  
			else (SELECT max(age) from ##mgw as b where b.rownumber<=a.act and a.gppraccode=b.gppraccode)+D*((select min(age) from ##mgw as b where b.rownumber>=a.act and a.gppraccode=b.gppraccode)-(select max(age) from ##mgw as b where b.rownumber<=a.act and a.gppraccode=b.gppraccode)) end as PercentileValue
	FROM ##mgw_percentileHolding as a
) as internal
group by gppraccode,N

Row_number(), dense Rank rank etc

admin — Wed, 02 May 2012 21:19:01 +0000

You need SQL 2005 or greater for this:
Rownumbers with partitions:

SELECT
    s.gppraccode
    ,age
    ,ROW_NUMBER()OVER (PARTITION BY gppraccode ORDER BY age asc) As [RowNumber]
FROM mgw_ane_base AS s

SELECT s.gppraccode, age
    ,ROW_NUMBER() OVER (ORDER BY age) AS "RowNumber"
    ,RANK() OVER (ORDER BY age) AS Rank
    ,DENSE_RANK() OVER (ORDER BY age) AS "DenseRank"
    ,NTILE(10) OVER (ORDER BY age) AS Quartile
INTO ##mgw
FROM mgw_ane_base AS s