You can either sync with master or wait until we do an official release. I’m pretty happy with how DataFrames are printed now.
Because there’s no substantive distinction between a DataFrame, a row of a DataFrame or part of a row, we print them the same way. I don’t really see why we should change that. Being more consistent with Base makes sense, but DataFrames aren’t arrays and I wouldn’t want people to think they are. The distict printing helps to reinforce an important conceptual distinction. We also just turned off printing out summaries because so many people have voiced complaints about it. I’m not sure we should reinstate it, since it causes confusion all the time. Even people who sort of understand the rules for printing didn’t really understand them: it seemed to be the kind of thing that only made sense to people who’d read the source code. — John On Jun 9, 2014, at 8:09 PM, cnbiz850 <[email protected]> wrote: > Thanks John for the update. Do I need to do a clone to get the new change? > > I think there are two issues here. One is to display df, a row of it, or > part of a row, like > > julia> df > julia> df[3, 1:20] > > The other is to print out (to the terminal or to a file). > > On the first issue, I hope DataFrames can adopt a similar behavior of arrays > (when the user tries to display an array that is large, julia prints out ... > in between the data). This is not only helpful to the user but is also > consistent with the overall julia style. > > On the second, if an entire df is to be printed to the terminal, perhaps it > should opt to print a summary only. If a row is to be printed, it should > always print out the data, however many elements. > > On 06/10/2014 10:43 AM, John Myles White wrote: >> Yup, this is the behavior that just got removed since we’ve had so many >> complaints. >> >> — John >> >> On Jun 9, 2014, at 7:42 PM, Taylor Maxwell <[email protected]> wrote: >> >>> If your dataframe has 6 or more columns then it prints this summary. If >>> you index less than 6 columns it will print ok. You can force it to spit >>> it all out with showall(df) but that may not be entirely legible. There >>> may be other ways to display it. >
