Thanks as always to Kit Baum, a new version of the descsave package is now available for download from SSC. In Stata, use the ssc command to do this, or adoupdate if you already have an old version of descsave.
The descsave package is described as below on my website. The new version fixes a speed issue. It has been streamlined to increase speed when using big (eg gigabyte-sized) input datasets, either on disk or in memory. (Let me know if it doesn't.)
I would like to thank Kit Baum for doing some informal tests on the new descsave with one of his own big datasets, and Alan Riley of StataCorp for alerting me personally to the speed issue at the 2022 UK Stata Conference. And also for whoever it was that reported the issue to Alan.
Best wishes
Roger
---------------------------------------------------------------------------------
package descsave from http://fmwww.bc.edu/RePEc/bocode/d
---------------------------------------------------------------------------------
TITLE
'DESCSAVE': module to export data set and machine-readable codebook
DESCRIPTION/AUTHOR(S)
descsave is an extension of describe, creating up to 2 output
data sets. These are a Stata data file with 1 observation per
variable and data on the descriptive attributes of each variable
(name, storage type, format value label, variable label, and
characteristics if specified), and also a do-file, which can be
called to reconstruct these descriptive attributes. The do-file
created by descsave can be used if the data set has been saved
using outsheet and input again using insheet. This can be useful
if the user wishes to create a definitive generic spreadsheet
version of the data, and to know that the original Stata version
can be reconstructed from the definitive generic version.
However, descsave can also be used when the user uses
parmest,label after a regression command, using xi dummy
variables for multilevel factors, and then reconstructs these
multilevel factors from the variable label in the parmest output
data set, in order to create tables and/or plots of confidence
intervals.
KW: data manipulation
Requires: Stata version 16.0
Distribution-Date: 20220927
Author: Roger Newson, King's College London
Support: email [email protected]
INSTALLATION FILES (click here to install)
descsave.ado
descsave.sthlp
---------------------------------------------------------------------------------
(click here to return to the previous screen)
The descsave package is described as below on my website. The new version fixes a speed issue. It has been streamlined to increase speed when using big (eg gigabyte-sized) input datasets, either on disk or in memory. (Let me know if it doesn't.)
I would like to thank Kit Baum for doing some informal tests on the new descsave with one of his own big datasets, and Alan Riley of StataCorp for alerting me personally to the speed issue at the 2022 UK Stata Conference. And also for whoever it was that reported the issue to Alan.
Best wishes
Roger
---------------------------------------------------------------------------------
package descsave from http://fmwww.bc.edu/RePEc/bocode/d
---------------------------------------------------------------------------------
TITLE
'DESCSAVE': module to export data set and machine-readable codebook
DESCRIPTION/AUTHOR(S)
descsave is an extension of describe, creating up to 2 output
data sets. These are a Stata data file with 1 observation per
variable and data on the descriptive attributes of each variable
(name, storage type, format value label, variable label, and
characteristics if specified), and also a do-file, which can be
called to reconstruct these descriptive attributes. The do-file
created by descsave can be used if the data set has been saved
using outsheet and input again using insheet. This can be useful
if the user wishes to create a definitive generic spreadsheet
version of the data, and to know that the original Stata version
can be reconstructed from the definitive generic version.
However, descsave can also be used when the user uses
parmest,label after a regression command, using xi dummy
variables for multilevel factors, and then reconstructs these
multilevel factors from the variable label in the parmest output
data set, in order to create tables and/or plots of confidence
intervals.
KW: data manipulation
Requires: Stata version 16.0
Distribution-Date: 20220927
Author: Roger Newson, King's College London
Support: email [email protected]
INSTALLATION FILES (click here to install)
descsave.ado
descsave.sthlp
---------------------------------------------------------------------------------
(click here to return to the previous screen)
