[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: Re: Unix stata big dataset

From   "Michael Blasnik" <>
To   <>
Subject   Re: st: Re: Unix stata big dataset
Date   Thu, 29 Nov 2007 19:29:17 -0500


It is still difficult to understand what you want or how to make it work faster.

1) You say you have 18 million observations -- is that for both datasets or just one dataset? How many observations are in the smaller dataset?

2) In your example data, do you expect to have 3 observations for id=1, or 9 observations (all combinations of the 3 observations in each dataset)? If you want three observations, how do you tell which observations to match from each dataset?

3) Do you want all observations for both datasets, or are there many observations in the larger dataset that don't match and you don't need?

Michael Blasnik

----- Original Message ----- From: <>
To: <>
Sent: Thursday, November 29, 2007 5:36 PM
Subject: Re: st: Re: Unix stata big dataset

The memory that I can allocate is 15 gb, the total observations are 18 millions

I have duplicate observation but I can't drop because they are the spell of work for each person, and I need these observations.
The two dataset have in common the same id, so I need to match the data set in booth but id is not unique.

so, one data set is
id x1 x3 x 4...
1 0 1991 1998
1 1 1991 1998
1 2 1999 1999
and second is:
id y1 y2 y3
1 34 2 35
1 34 2 67
1 34 1 68

the idea is to keep all the people that have the same id to obtainer this data set
id xi x2 x3 y1 y2 y3

Sorry I don't understand the last part of email , how to do the merge
thanks a lot for your help
*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index