Even within a single data source, drink names are not necessarily unique. For Wikidata, one such example is the drink "Mudder's Milk" (Q89590905 and Q93878948). Not only does this introduce nasty UI bugs if the name is assumed to be unique (e.g., when used as the key attribute in react) but also renders the design of drink-dependent APIs difficult. We should thus introduce custom identifiers, especially if our data set is supposed to be robust against changes in external data sets.
In addition to the new identifier, we should store data source-dependent identifiers. This may even allow us to merge the information available on multiple sources or to avoid duplicates at worst.
Even within a single data source, drink names are not necessarily unique. For Wikidata, one such example is the drink "Mudder's Milk" (Q89590905 and Q93878948). Not only does this introduce nasty UI bugs if the name is assumed to be unique (e.g., when used as the
keyattribute in react) but also renders the design of drink-dependent APIs difficult. We should thus introduce custom identifiers, especially if our data set is supposed to be robust against changes in external data sets.In addition to the new identifier, we should store data source-dependent identifiers. This may even allow us to merge the information available on multiple sources or to avoid duplicates at worst.