* fix #170 persist and coalesce before to local iterator while using incremental result collect * add log * use df.rdd.getPartitionNum as upper limit * log & ut * bypass coalesce if df.rdd.partition.size is small * add ut * code cov * code cov |
||
|---|---|---|
| .. | ||
| if | ||
| src | ||
| pom.xml | ||