I get no result with lazy-generator: https://groups.google.com/d/topic/cascalog-user/4blpXGKqSkI/discussion

i_get_output.clj

;; this one outputs to stdout until it dies due to 'Exceeded max jobconf'

(defn format-package-repo-for-cassandra
  [source-type base-attributes sink repo]
  (cio/with-fs-tmp [_ tmp]
    (let [packages (package-seq repo)]
      (if-not (empty? packages)
        (let [lazy-tap (ops/lazy-generator tmp packages)]
          (?<- (stdout)
               [?row-key ?col ?val]
               ((vec packages) :> ?name _ _ _ _ ?attr)
               (format-package-for-output :< source-type base-attributes ?attr :> ?row-key ?col ?val)))))))

no_output.clj

;; I get no output to stdout

(defn format-package-repo-for-cassandra
  [source-type base-attributes sink repo]
  (cio/with-fs-tmp [_ tmp]
    (let [packages (package-seq repo)]
      (if-not (empty? packages)
        (let [lazy-tap (ops/lazy-generator tmp packages)]
          (?<- (stdout)
               [?row-key ?col ?val]
               (lazy-tap :> ?name _ _ _ _ ?attr)
               (format-package-for-output :< source-type base-attributes ?attr :> ?row-key ?col ?val)))))))

This turned out to be a bug in 1.8.6, it's fixed in 1.8.7. lazy-generator wouldn't work with nullable vars.