Analysis

an error is created if a vl already exist.

add new option , e.g.
new vl g i (0,"No") (1,"possibly") (2,"Definetely") !replace;

merged datasets get difficult names:

merge !ds:=ds5;

list ds _m*;

// example _merge_ds1_ds5_1 _merge_ds1_ds5_2 etc

Simplify "_merge_" to: "m-"

Label of the merged dataset is now the same as the first(use) dataset.

- change to: "("+"label of first"+"+"(" + "label of second ds"+")"

export to stata

sdato,adato,atime

.,25-06-2017,12:12:00

.,20-12-2017,10:38:00

will be exported to Stata 14:

sdato adato atime

2.15e+09 25jun2017 12:12:00

2.15e+09 20dec2017 10:38:00

date is exported as 2.15e+09, not as the Stata "." value

Two aspects of modifications of value labels are not handled in a user oriented manner.

1 change to "not found": // value label should be forgiving if deleting a value is not found, just like "drop v " .

for S in ["alvorliglbl", "fvgl","janej","henvislbl","koenlbl","plbl"] do

edit vl @{s} !d:=1410065407;

ERROR: Cannot delete - value does not exist!

2 change to "Cannot Assign Missing to undefined Value in: xxxx (here alderlbl)"

// Assigning missing to an undefined value is not handled correctly

edit vl alderlbl !m:=9;

ERROR: Unhandled error during execution:

Access violation

Please contact EpiData (info@epidata.dk) and include a copy of history and if possible example data!

.Append !fn:="Lykke.epx" ;
ERROR: This file is locked by another program!
You can force a read/save by using the option !force

.Append !fn:="Lykke.epx" !force ;
ERROR: Unknown option/variable: force

Create "engine for graphs" - much like the aggregate/table system. That is a common class for all graphs which may then be inherited by as well SPC as standard graphs, unless SPC need their own class.

Total functionality of graphs defined as in EpiData Classic Analysis

* EPIDATA ANALYSIS PROGRAM

logopen "H:\Teaching Courses\Outbreak Course\bugTN_06122013\bugTN_06122013.htm" /close /replace

cd "H:\Teaching Courses\Outbreak Course\bugTN_06122013"
read "H:\Teaching Courses\Outbreak Course\bugTN_06122013\wedd" /close

set random simulations=10000

***********************************************
* EpiData Analysis 2.2.1.171
* tables ill redsauce gheerchick /o /nt /nc /ex /t /obs
* Using 2.2.1.171: OR = 1.08 [95% CI 0.22 to 5.66]; p(M-H)=0.916; p(exact)=0.0414

* EpiData Analysis 2.2.2.182
tables ill redsauce gheerchick /o /nt /nc /ex /t /obs /oex
* Using 2.2.2.182: OR(Exact) = 1.08 [95% CI 0.22 to 5.66]; p(M-H)=0.046; p(exact)=0.0414
* Using 2.2.2.182: OR = 1.07 [95% CI 0.11 to 10.06]

* Epi Info for DOS v6.04d
* tables redsauce ill gheerchick
* Using Epi Info for DOS v6.04d: OR (MLE) = 1.08 [95% CI 0.22 to 5.66]; p(M-H uncorrected)=0.91644269; p(M-H corrected)=0.81105112

* Stata/SE 12.1 for Windows (32 bit) Revision 20 Mar 2013
* logit ill redsauce gheerchick
* P-value for coefficient of redsauce = 0.916

* The estimate is close to 1 and the confidence interval is wide, so a test for the adjusted OR being significantly different to 1 should have a p-value close to 1
* Both Epi Info for DOS and Stata give a p-value of 0.916
* So I have a lot of confidence in this being the correct p-value.

* EpiData Analysis (build 171)
* The chi-squared test statistic and the p-value for a M-H weighted odds ratio using the option /t is 0.01 and 0.916
* This p-value using the option /t appears to be correct.
* The p-value using the option /ex appears to be incorrect.
* The chi-squared test statistic and the p-value using the option /ex is 6.147 and 0.0414.
* If the test statistic is correct a comparison to a chi-squared distribution of 1 degree of freedom should give a p-value equal to 0.0132?
* Possibly the correct p-value should be calculated as 1-2*(0.0414) = 0.917?

* EpiData Analysis (build 182)
* The chi-squared test statistic and the p-value for a M-H weighted odds ratio using the option /t is 6.15 and 0.046
* This p-value using the option /t appears to be incorrect.
* The p-value using the option /ex appears to be incorrect.
* The chi-squared test statistic and p-value using the option /ex is 6.147 and 0.0414.
* If the test statistic is correct a comparison to a chi-squared distribution of 1 degree of freedom should give a p-value equal to 0.0132?
* Possibly the correct p-value should be calculated as 1-2*(0.0414) = 0.917?

***********************************************
logclose

using sked.rec:

missingvalue chicken /8
missingvalue turkey /8
* Expected to get odds ratios in the following command (has the output from ct changed from * previous builds? - if so, not listed as a change):
tables status sex beef chicken turkey /sa /ct /t
* Need to specify o option:
tables status sex beef chicken turkey /sa /ct /t /o
* But odds ratios disappear if add /ex (with or without t option):
tables status sex beef chicken turkey /sa /ct /t /o /ex

Det ville være dejligt hvis man kunne kopiere en enkelt linje fra sin historik til pgm filen.
Tak.
Mvh Anna

Using sked.rec:

read sked.rec
missingvalue turkey /8
missingvalue chicken /8
tables status beef turkey chicken /ct /oex /t /sa

yields the MLE ORs all equal to 0.00 and CIs 0.00-0.00

I noticed that the GUI which allows you to create histograms (Graph -> Histogram) does not let you specify a boolean variable as "by" variable.

Example:

. read C:\EpiData\TestData\bromar.rec
. select dectime<>.
. gen b fast=(dectime<3.87)

it is possible to do the following:

. his age /by=fast

but the GUI form does not list "fast" as a "by" variable.

Ver 2.2.1.171

Helpfile says for the command "!":

dos textexecute any valid MS-DOS command and return to EpiData- dos command will open an MS-DOS window/open : Keep window open after execution
!works only on XP+ Pc's

I interpret this as working with XP or higher, but if it works in some XP versions (perhaps), it does not do so in all: I know of two colleagues in whom the command to open a batch file that directs to open a REC file (in EpiData Entry) opens the REC file in the default text editor.

Vista and Windows 7 no problem, it works perfectly.

Any chance to get this command "!" to work on XP machines? It is a marvelously powerful command. Or is there a workaround? The reason for its usefulness comes with using EpiData Analysis for a menu report, and allowing (or rather directing) people to enter data into a predetermined file only, and then in the same menu allow analysis by button pressing - the whole package including software can be given to people who need to know nothing.

Example is on www.tbrieder.org | EpiData | Part D, Exercise 4 solution

The use of freq * /m instead of freq * /M troncate the display of results-> the result list is cut.

Strange problem. When I run this command on the attached data set everything runs fine: "stab komplians tjekind tjekud timeout /stat="sum" /by="yearweek" /n".

However, running this command freezes EpiData: "stab tjekind1 tjekind2 /stat="sum" /by="yearweek" /n" EpiData apparently goes into some kind of infinite loop using up all the cpu power. I can force the program to close by clicking the cross in the window corner.

The problem appears whenenver there are variables with almost similar names, e.g. tjekud1 tjekud2, etc.

With these data and this command the output table does not mark test 5 as positive. Tests 4 and 5 should both be positive. The graph looks correct to me.

### command line and data table ###
ichart reghn uge /t4 /t5 /sl

Uge RegHN
2009-15 185
2009-16 272
2009-17 310
2009-18 241
2009-19 197
2009-20 277
2009-21 282
2009-22 316
2009-23 237
2009-24 302
2009-25 277
2009-26 267
2009-27 240
2009-28 261
2009-29 228
2009-30 167
2009-31 183
2009-32 181
2009-33 204
2009-34 215
2009-35 249
2009-36 291
2009-37 332
2009-38 303
2009-39 305
2009-40 292

b170:

Using sked.rec:

BOXPLOT age /BY=SEX /legend

gives Age next to both colours in the legend box. I might have thought it should have showed Male Female against the respective colours?

The search and find dialog is buggy and requires a rewrite.

This is scheduled to happen after the new core principle has been implemented.

A completely new editor is needed. The current is buggy and requires to much rewriting to maintain at a reasonable level.

This is expected to be implemented after the new core principle has been implemented.

There has been a few comments on a "Dummyform already exists" error when running graph command.

This is not yet confirmed.

a: Implement Masks on Breaks edit and Freeze edit.
b: Add labels for Title, subtitle etc. on the "Titles" page.

Analysis v 2.2 (build 169 and earlier)

read /cb failed with import of a single column or if there were any missing values in the last numeric column imported
Workaround: create a new column in Excel that had no missing values or include . for missing values

Reads OK:
test note n2
0.025974026 1 a
0.019933555 1 b
0.013333333 1 c
0.049382716 1 d
0.034059946 1

reads OK:
test note n2
0.025974026 1 a
0.019933555 1 b
0.013333333 1 c
0.049382716 1 d
0.034059946 e

does not read:
test note n2
0.025974026 1 a
0.019933555 1 b
0.013333333 1 c
0.049382716 1 d
0.034059946

error message:
Separators: (tab: 10)(semicolon: 0)(comma: 0)(space: 0) Lines: 6
Cannot define number of variables

We need a general document about where we use the /W option and how a new command should be implemented with this option.