tbl_df example is now broken for Lahman 3.0

[friendly]

v 3.0-1 of the Lahman package entailed some changes in the names of ID variables:

• HallOfFame$hofID is now HallOfFame$playerID
• managerID is now playerID in all tables
• Removed from Master: managerID, hofID, holtzID, lahmanID, lahman40ID, lahman45ID, nameNote, nameNick, and college

Consequently, when submitted to CRAN, the following error is generated when checking dplyr:

< > ### Name: tbl_df
< > ### Title: Create a data frame tble.
....
< > data("Master", "HallOfFame", package = "Lahman")
< > player_info <- select(tbl_df(Master), playerID, hofID, birthYear)
< Error in eval(expr, envir, enclos) : object 'hofID' not found
< Calls: select ... select_vars -> select_vars_q -> lapply -> FUN -> eval

The fix is to remove hofID from the call to select on line 53 of tbl_df.Rd

-Michael

[hadley]

I'm travelling at the moment, but I'll get fixed as soon as I can.

[hadley]

Removing that variable fixes the first problem, but then the joins won't work with the current lahman release :/

[friendly]

Well, we're in a package-version-clash situation here. The CRAN people rejected the new version, on R-Forge
because of the problem with dplyr, and my suggested fix won't work with the old version. I'll try asking the CRAN people to approve the new Lahman package as is, and then you can update dplyr, unless you have any other suggestion.

[friendly]

Lahman 3.0-1 is on CRAN now and should propagate shortly. Can you install the new version
and try the example again?
thx,

[djmurphy420]

Two changes are needed:

(i) remove hofID in the call that produces player_info
(ii) in the next line, replace hofID with playerID.

When I made those two changes, all of the joins worked in the latest version of the Lahman package. I'm using version 0.2 of dplyr for this test. I tried to do the same with the development version of dplyr but it failed to install...that's a separate and tangential issue, though.