s-news
[Top] [All Lists]

CRSP data and big data time series

To: <s-news@lists.biostat.wustl.edu>
Subject: CRSP data and big data time series
From: "Chen,Sichong" <sichong.chen@postgrad.manchester.ac.uk>
Date: Wed, 28 Jun 2006 01:30:10 +0800
Thread-index: AcaaD1dG9itXE1j5QUWtb/hW1mUkCg==

Dear users:

 

Sorry for receiving this mail twice

 

I have two questions regarding manipulation of the data obtained from CRSP and converted to big data time series object as I am new to CRSP data and big data object.

 

My data imported as big data object US is:

PERMNO

DATE

COMNAM

PRC

VOL

10,001.00

19,950,103.00

ENERGY WEST INC

8.38

372

10,001.00

19,950,104.00

ENERGY WEST INC

8

8,275.00

10,002.00

19,950,103.00

SOUTH ALABAMA BANCORPORATION INC

13.32

0

10,002.00

19,950,104.00

SOUTH ALABAMA BANCORPORATION INC

14.12

0

10,002.00

19,950,105.00

SOUTH ALABAMA BANCORPORATION INC

13.17

0

10,003.00

19,950,103.00

GREAT COUNTRY BK ASONIA CT

2.32

0

10,003.00

19,950,104.00

GREAT COUNTRY BK ASONIA CT

2.12

0

10,003.00

19,950,105.00

GREAT COUNTRY BK ASONIA CT

2.48

5,000.00

 

First question is that I use y=bdTimeSeries(US[,c(1,3,4,5)],positions.=US[,2]), I try to convert this object into big data time series data set. But strangely, I can not find object y in my object explorer window, but print(y) works and it shows that y is a time series object as traditional time series object shows. But why I can not find this object in object explorer window?

 

The second question is that, I want have two matrix, price and trading volume:

Price:

PERMNO

10,001.00

10,002.00

10,003.00

19,950,103.00

8.38

13.32

2.32

19,950,104.00

8

14.12

2.12

19,950,105.00

          NA

13.17

2.48


Trading volume:

PERMNO

10,001.00

10,002.00

10,003.00

19,950,103.00

372

0

0

19,950,104.00

8,275.00

0

0

19,950,105.00

         NA

0

5,000.00

 

 

 

 

The purpose to have two matrix is to get a new return matrix by using the price matrix. Then divide this new return matrix by the trading volume matrix by matching the time period. I am not sure whether it is right to split CRSP dataset into two matrix to do want I want to do and whether it will save more time. But this is the only way I figure out to get what I want.

 

Any suggestions or new ways to deal with this are appreciated.

 

Many thanks,

Sichong

SPlus 7.0 & FinMetrics 2.0 User

Windows XP

 

<Prev in Thread] Current Thread [Next in Thread>