[
https://issues.apache.org/jira/browse/DRILL-3249?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16711800#comment-16711800
]
Vitalii Diravka commented on DRILL-3249:
----------------------------------------
Looks like it is no more an issue:
{code}
vitalii@vitalii-pc:/tmp$ cat /tmp/empty_csv/0_0_0.csv
vitalii@vitalii-pc:/tmp$
{code}
{code}
0: jdbc:drill:zk=local> CREATE TABLE dfs.tmp.empty_parquet_table(dsa_id, uid,
DN, d_csi_name, d_svc_avail, d_sending_opt, d_supp_cap_ph, d_repl_hndl_lup,
d_repl_hnd_interr, d_opt_routing, ref_status) AS
. . . . . . .semicolon> SELECT cast(columns[0] AS INTEGER), columns[1],
columns[2], columns[3], columns[4], columns[5], columns[6], columns[7],
columns[8], columns[9], columns[10] FROM dfs.`/tmp/empty_csv/0_0_0.csv`;
+-----------+----------------------------+
| Fragment | Number of records written |
+-----------+----------------------------+
| 0_0 | 1 |
+-----------+----------------------------+
1 row selected (0.416 seconds)
0: jdbc:drill:zk=local> select * from dfs.tmp.empty_parquet_table;
+---------+-------+-------+-------------+--------------+----------------+----------------+------------------+--------------------+----------------+-------------+
| dsa_id | uid | DN | d_csi_name | d_svc_avail | d_sending_opt |
d_supp_cap_ph | d_repl_hndl_lup | d_repl_hnd_interr | d_opt_routing |
ref_status |
+---------+-------+-------+-------------+--------------+----------------+----------------+------------------+--------------------+----------------+-------------+
| null | null | null | null | null | null | null
| null | null | null | null
|
+---------+-------+-------+-------------+--------------+----------------+----------------+------------------+--------------------+----------------+-------------+
1 row selected (0.145 seconds)
{code}
> CTAS from an empty TSV file fails
> ---------------------------------
>
> Key: DRILL-3249
> URL: https://issues.apache.org/jira/browse/DRILL-3249
> Project: Apache Drill
> Issue Type: Bug
> Components: Storage - Text & CSV
> Affects Versions: 1.0.0
> Reporter: Stuart Hayes
> Priority: Major
> Labels: csv
> Fix For: 1.14.0
>
>
> Performing a CTAS to parquet from an empty CSV file fails:
> CREATE TABLE dfs.latest.d_csi(dsa_id, uid, DN, d_csi_name, d_svc_avail,
> d_sending_opt, d_supp_cap_ph, d_repl_hndl_lup, d_repl_hnd_interr,
> d_opt_routing, ref_status) AS
> SELECT cast(columns[0] AS INTEGER), columns[1], columns[2], columns[3],
> columns[4], columns[5], columns[6], columns[7], columns[8], columns[9],
> columns[10] FROM dfs.converted.`02062015/25/d_csi.txt`;
> Error: SYSTEM ERROR: java.lang.IllegalArgumentException: MinorFragmentId 0
> has no read entries assigned
> This is part of a daily automated job in which case the entire job fails. An
> empty file/table should be a valid scenario.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
