I am currently working on a rather large cross-sectional data set.
Overall I have data comprising around half a Million individuals, about
225000 households and 26 countries.
I tried to merge the different datasets, after dropping irrelevant
variables before. But merging households (identified by HID) and
countries (country) by "join" really takes a lot of time. Actually,
joining one variable takes 25 minutes on my linux machine (2.6GHz). If I
use STATA it may take a minute or so.
Why does it take that long here? I am surprised because typically gretl
operates pretty fast.
Also, is there a way to merge all cross-sectional variables from the
"outside" dataset with the "inner" one by a single command? At the
moment one has to specify a join command for each variable separately,
right? I am just asking out of curiosity as I am fine with the way it is
A heads-up for those using gretl's "join" command: there are some changes
in current CVS and the snapshots for Windows and Mac. Most of the changes
represent substantial enhancements. In particular we now support a much
more flexible filter; see the updated chapter on join in the current Gretl
User's Guide for details.
But please note a couple of backward-incompatible points:
1) We now check the column headings in the CSV file from which data are to
be drawn, and if they're not valid gretl identifiers (e.g. they contain
spaces) we "fix" them. In the context of the join command you must use the
fixed names. The fix-up rule is quite simple and is set out in the help
for the new fixname() function, which you can use if you're in doubt as to
what will happen to a given column name under the new policy. Example:
eval fixname("1. Some name")
gives the result "Some_name".
2) If you want to use a string constant in a join filter you must enclose
it in backslash-escaped double quotes. For example, suppose the CSV file
contains a string-valued column named "gender" with values "F" and "M" and
you want to filter on gender = "F". Your filter option must then look like
Alternatively you can define a string variable and use it in place of the
string constant. So after doing
string wanted = "F"
you could use
and get the same result.
Sorry to raise an issue which has been raised before (see http://lists.wfu.edu/pipermail/gretl-users/2013-January/008426.html)
If I take a csv file
Then import in gretl (v1.9.12 build date 2013-13-15) gretl still fails to recognise a timeseries! :
using delimiter ','
longest line: 16 characters
first field: 'idate'
number of columns = 2
number of variables: 2
number of non-blank lines: 3
scanning for variable names...
scanning for row labels and data...
treating these as undated data
note it also re-labels the date column as idate
Dear Gretl users,
I have a spreadsheet file (simulation) which is attached , I'm trying to
import it as panel data but if I try the built in import command from GUI I
only can import the first sheet.
Can you give me any suggestion?
I obtain an error running the following script + panel dataset. The
error message I receive is
*** Error in `/usr/bin/gretl_x11': corrupted double-linked list:
and gretl stops working. I am using current cvs on ubuntu.
I played a bit around with your nice HT-IV package.
I obtain a segmentation fault ("Speicherzugriffsfehler" in German) if
one/some of the Z1 (TIV exog.) variables is/are a zero series. The error
Deleting var 50 (FR) at list pos 10: all zero
Deleting var 49 (ES) at list pos 9: all zero
Speicherzugriffsfehler (Speicherabzug geschrieben)
I tried to replicate the error using your example script by adding a
zero series to the list, but I couldn't replicate it. However, I obtain
it using my own current script.
Dear Gretl community,
I'm using Gretl on a data-set and I'm applying 'hsk' function.
I read it is useful when heteroskedasticity can be approximated by a quadratic relationship with the regressors.
I'd like to know if there are some methods in the programs to evaluate this feature and even if there are some other hypothesis behind the 'hsk' function.
Thank you very much,
I am trying to replicate Baltagi's example of the HT-IV. I wrote the
following small function, but do not obtain exactly his results (see
Actually, I am not so sure whether I've used the correct (transformed)
Maybe somebody of you has already dealt with this, and may help me out.
open Baltagi_schooling_wage.gdt --quiet
series EXP2 = EXP_^2
# List definition
list X1 = OCC_ SOUTH_ SMSA_ IND_ #TV exogeneous
list Z1 = FEM_ BLK_ #Time-invariant exogeneous
list X2 = EXP_ EXP2 WKS_ MS_ UNION_ #TV endogenoeus
list Z2 = ED_ #TIV endogenoeus
function void xthtaylor (series Y "dependent",
list X1 "TV exog.",
list Z1 "TIV exog.",
list X2 "TV endog.",
list Z2 "TIV endog.")
# X2(i,t) are instrumented by their deviation from individual means
list X2inst = null
loop foreach i X2 -q
series dev_$i =X2.$i - pmean(X2.$i)
X2inst += dev_$i
# Z2(i) are instrumented by the individual avg of X1(i,t)
list Z2inst = null
loop foreach i X1 -q
series m_$i = pmean(X1.$i)
Z2inst += m_$i
#TSLS - Hausman-Taylor (HT)
tsls Y const X1 X2 Z1 Z2 ; const X1 X2inst Z1 Z2inst --no-tests
I attached the config.log
On Fri, Sep 6, 2013 at 4:27 PM, cociuba mihai <cociuba(a)gmail.com> wrote:
> the config.log is attached
> On Fri, Sep 6, 2013 at 3:02 PM, cociuba mihai <cociuba(a)gmail.com> wrote:
>> Dear Gretl users,
>> I'm trying to install Gretl from cvs in Ubuntu 13.04, 64b,
>> at the ./configure --enable-build-doc command I receive the error
>> Please install lapack and then reconfigure gretl.
>> Lapack is available from http://www.netlib.org/lapack/
>> but the liblapack-dev is already installed.
>> You can find the output from the configure command attached.
>> Any suggestions?