Some querying tweaks with good performance improvement

[MikeYalter]

Doing a query vs solr with q=(*:*) AND (field:"test") is a lot slower than q=(field:"test"). Something close to 50% speed improvement on ad hoc tests for field queries.

[markwoodhall]

Hi Mike,

I've been doing a little bit of testing against this branch using some of the automated tests created recently. It looks like I'm hitting a problem when trying to load a DOI using something like http://localhost:3000/works/10.13039/501100003593, this works fine on master but using this branch I get a NullPointerException with the following stack trace:

"stack": [
      "clojure.lang.Numbers.ops(Numbers.java:1013)",
      "clojure.lang.Numbers.gt(Numbers.java:229)",
      "clojure.lang.Numbers.gt(Numbers.java:3864)",
      "cayenne.api.v1.query$__GT_solr_query.invokeStatic(query.clj:331)",
      "cayenne.api.v1.query$__GT_solr_query.doInvoke(query.clj:264)",
      "clojure.lang.RestFn.invoke(RestFn.java:439)",
      "cayenne.data.work$fetch_one.invokeStatic(work.clj:174)",
      "cayenne.data.work$fetch_one.invoke(work.clj:170)",
      "cayenne.api.v1.routes$work_resource$fn__23538$fn__23539.invoke(routes.clj:222)",
      "cayenne.api.v1.routes$__GT_1$fn__23471.invoke(routes.clj:66)",
      "liberator.core$decide.invokeStatic(core.clj:98)",
      "liberator.core$decide.invoke(core.clj:91)",
      "liberator.core$exists_QMARK_.invokeStatic(core.clj:406)",
      "liberator.core$exists_QMARK_.invoke(core.clj:406)",
      "liberator.core$decide.invokeStatic(core.clj:103)",
      "liberator.core$decide.invoke(core.clj:91)",
      "liberator.core$processable_QMARK_.invokeStatic(core.clj:409)",
      "liberator.core$processable_QMARK_.invoke(core.clj:409)",
      "liberator.core$decide.invokeStatic(core.clj:103)",
      "liberator.core$decide.invoke(core.clj:91)",
      "liberator.core$encoding_available_QMARK_.invokeStatic(core.clj:413)",
      "liberator.core$encoding_available_QMARK_.invoke(core.clj:413)"

A quick glance seems to indicate (.getRows query) might be returning nil.

[MikeYalter]

Hi Mark, yeah, I actually had a check in place with the (and) but I must have hit undo too many times in the ide at some point.

[markwoodhall]

Looks like it works fine now Mike, thanks. Do you happen to know if there is any implication to running (.getRows query) twice, like we do now?

If there is, perhaps we could use something like (> (or (.getRows query) 0) 0)? If not then I will merge this.

[MikeYalter]

@markwoodhall there shouldn't be as it's generated from a map that we created, rather than a return from the query. But I like your way of doing the check.