The RATS Software Forum

Posted: **Thu Jun 25, 2015 7:58 am**

Hello Tom,
I have an excel file with five day a week data staring from 1-1-2001 to 31-12-2014. I write the following command to load the data into RATS for analysis. The command that I used is

Code: Select all

OPEN DATA "C:\Users\Muhammad Irfan Awan\Desktop\Exchange_KSE.xlsx"
CALENDAR(D) 2001:1:1
DATA(FORMAT=XLSX,ORG=COLUMNS,SHEET="KSE_SBP_5DAYS") 2001:01:01 2014:12:31 date1 kse sbp

Then find the basic statistics for this data by
stat kse
I got the following output
Statistics on Series KSE
Daily(5) Data From 2001:01:01 To 2014:12:12
Observations 177 Skipped/Missing 3463
Sample Mean 10197.801638 Variance 51806776.049777
Standard Error 7197.692411 SE of Sample Mean 541.011691
t-Statistic (Mean=0) 18.849503 Signif Level (Mean=0) 0.000000
Skewness 0.960243 Signif Level (Sk=0) 0.000000
Kurtosis (excess) 0.706196 Signif Level (Ku=0) 0.060099
Jarque-Bera 30.878956 Signif Level (JB=0) 0.00000
What went wrong with my commands to that it give 3463 missing observations?

Another question is Is it possible to load data from excel sheet with unbalance dated data? The attached file contains the data in sheet

KSE_SBP_SAMEDAy

of irregular dates. The sheet

KSE_SBP_5DAYS

contains data with 5 days a week.

Please guide me so that I can upload my data

Thanks
Irfan

Posted: **Thu Jun 25, 2015 10:16 am**

Your date fields are an odd collection of different styles. The first ten are numeric codings with date formats, the next thirteen are strings, then numeric codings again, etc. RATS can handle a consistent choice of one or the other, but not a single column that switches between the two.

If you're trying to skip the missing entries, read the data as "irregular" time series rather than "daily". Daily means 5-day-a-week---if Friday is missing, an NA is inserted to keep the time sequence.

Posted: **Thu Jun 25, 2015 6:22 pm**

Dear Tom
I have same data in STATA format but same problem arise. The data file is attached with this post.
When I execute the command the following error on output window.

## M4. A memory request for an additional -168032 bytes cannot be satisfied

What should I do in this regards, Please guide me

Regards
Irfan

Posted: **Sun Jun 28, 2015 1:28 am**

Because the dates aren't coded as expected, you need to override the format with the DATEFORMAT option:

open data kse_ex_5days.dta
cal(d) 2001:1:1
data(format=dta,dateformat="mm-dd-yyyy") / date1 kse rate

It also reads fine if you take the cal(d) instruction out so it read as undated data.

Posted: **Wed Aug 26, 2015 11:48 am**

Dear Tom,
I want to graph this series but the dates of the data are 5 days a week but some working days are missing when I upload this file into RATS it give NA in place of missing observation on a given date. What should I do to avoid this problem also it did not read the data till the end. The data file is attached with this post. I have use the command mentioned in your previous reply to me.

Code: Select all

[ENTRY           KSE
 2001:01:02   1522.160034180
 2001:01:03   1550.420043945
 2001:01:04   1550.040039063
 2001:01:05   1545.900024414
 2001:01:08   1522.040039063
 2001:01:09   1528.310058594
 2001:01:10   1529.510009766
 2001:01:11   1504.800048828
 2001:01:12   1470.719970703
 2001:01:15        NA
 2001:01:16   1465.229980469

Posted: **Wed Aug 26, 2015 1:48 pm**

Doesn't that graph OK? It should just put a thin dash to connect lines where there would be missing values. Given the amount of data, those probably are barely visible.

The RATS Software Forum

Reading of Unbalance dated data

Reading of Unbalance dated data

Re: Reading of Unbalance dated data

Re: Reading of Unbalance dated data

Re: Reading of Unbalance dated data

Re: Reading of Unbalance dated data

Re: Reading of Unbalance dated data