The program pwcorr2 displays and stores in a matrix "M_pwcorr2" (or exports to Excel) the first column of pwcorr. It is helpful if you have different sets of variables in your rows and columns If the variables in rows and columns are identical, then the results are the same as in the standard pwcorr.
It is somewhat similar to ssc package "cpcorr", but in addition handles missing variables and reproduces pwcorr output.
Code in "example.do" sets up a quick illustrative case. Where pwcorr and pwcorr2 produce the following output:
pwcorr columnvar1 columnvar2 rowvar1 rowvar2 rowvar3
column~1 column~2 rowvar1 rowvar2 rowvar3
-------------+---------------------------------------------
columnvar1 | 1.0000
columnvar2 | 0.0580 1.0000
rowvar1 | 0.5498 0.3301 1.0000
rowvar2 | -0.2973 0.3772 -0.0285 1.0000
rowvar3 | . . . . .
pwcorr2 "columnvar1 columnvar2" "rowvar1 rowvar2 rowvar3"
M_pwcorr2[3,2]
columnvar1 columnvar2
rowvar1 .54980936 .33012319
rowvar2 -.29734494 .37717309
rowvar3 . .