cloneDeep performance

[elado]

Compared to other clone packages or a simple JSON.parse(JSON.stringify(o)), _.cloneDeep is extremely slow. While JSON isn't always a good fit (e.g. when copying a function) it's still faster.

Is there anything _.cloneDeep is doing that the other implementation doesn't?

[jdalton]

Hi @elado!

Our deep clone perf compared to others or JSON will be slower. Ours tackles array buffers, buffers, dates, maps, nulls, regexes, sets, symbols, typed arrays, boolean/number/string objects, and circular references. If you know the data you're going to clone is JSON compatible then go for it. However, knowing if your data is JSON compatible isn't always straight forward
(See npm/pull/11306).

I'd like to avoid explicitly calling out perf. We already cover the things cloneDeep does and that it's recursive which is a good indicator.

[runvnc]

I think you should call out performance explicitly while at the same time mentioning that other cloning modules may not handle all of the cases you do.

cloneDeep performance issues have become the biggest slowdown in my current application so I will have to find another module. The reason I had used this clone rather than other modules, and also had not considered it was giving a problem was because of the reputation of lodash.

[jdalton]

@runvnc

cloneDeep performance issues have become the biggest slowdown in my current application so I will have to find another module.

Use the tool that's best for your situation. If a generic recursive clone with circular reference guards doesn't fit your needs, then by all means use another clone that does.