Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: -hcavar- "too many leaves"

From   Mark Taylor <>
To   "" <>
Subject   st: -hcavar- "too many leaves"
Date   Mon, 28 Oct 2013 10:22:13 +0000

Dear Statalist

Using Stata 13.0 for Mac, I'm using the user-written command -hcavar-. This command's available at, and the details on my machine are as follows

*! Version 3.2 15 April 2010

*! Jean-Benoit Hardouin

I'm using this command for hierarchical cluster analysis on a large number of variables, along the lines of


hcavar var1 var2 var3... var99,partition(1/30)


However, when I exceed 100 variables, I get the following output



Number of individuals with none missing values: 71577

Maximal number of modalities for a variable: 9

Proximity measures: Pearson

Method to aggregate clusters: Unweighted Pair-Group Method of Average


too many leaves; consider using the cutvalue() or cutnumber() options



In Googling this error message (such as at, it seems  that this error comes because the dendogram can't handle that many leaves - but I can't figure out how to turn off the dendogram. The help file says that {cmdnodendogram} enables the displaying of the dendogram, but typing -nodendogram-  tells me that option isn't allowed. (There are other methods to aggregate clusters which don't have this problem, such as -median-, but these aren't as useful for me.)

Any advice would be gratefully received.

*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index