Guava is a good library, but it has had some versioning issues in the past. I don't see why we shouldn't just use the hadoop util version or a stringbuilder, as appropriate.
Best, Colin On Dec 22, 2015 1:56 PM, "dam6923 ." <dam6...@gmail.com> wrote: > Hello! > > I have been looking through some of the Hadoop code and I see that > there are (at least) four different ways used to join a collection of > Strings with a delimiter: > > 1) org.apache.hadoop.util.StringUtils.join(...) > 2) org.apache.commons.lang.StringUtils.join(...) > 3) com.google.common.base.Joiner > 4) Manual - StringBuilder/For-Loop > > This question came to me when I was looking to replace some instances > of "Manual" with a library approach. Is there a preferred way of > doing it? The Hadoop StringUtils did not have the signature I was > looking for: StringUtils.join(Object[]). However, Apache Commons and > Google Guava both support the signature. > > I imagine one would want to "Eat their own dog food" and use the > Apache Commons library, however, there would be some issues on the > import list, trying to user two classes with the same name. Should I > bring over the required Join methods from Apache Commons and place > them into the Hadoop library? Should we be using the Joiner? > > Thanks. >
