[ https://issues.apache.org/jira/browse/SPARK-9807?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14712577#comment-14712577 ]
Yanbo Liang commented on SPARK-9807: ------------------------------------ This is not a bug. {map(lambda l: re.split(col_delimiter, l)} returns list of string and then `pyspark.sqlContext.createDataFrame` will convert the parsed lines to a PySpark DataFrame with all the columns are string type. If you want to make it a DataFrame with correct schema, you need to specify it like: {code} schema = StructType([ StructField("name", StringType(), True), StructField("age", IntegerType(), True)]) df3 = sqlContext.createDataFrame(rdd, schema) {code} > pyspark.sql.createDataFrame does not infer data type of parsed TSV > ------------------------------------------------------------------ > > Key: SPARK-9807 > URL: https://issues.apache.org/jira/browse/SPARK-9807 > Project: Spark > Issue Type: Bug > Components: PySpark > Affects Versions: 1.4.1 > Environment: CentOS 6, Python version 2.7.10, Scala version 2-10 > Reporter: Karen Yin-Yee Ng > Original Estimate: 24h > Remaining Estimate: 24h > > I tried parsing a space-separated file from HDFS. > And using `pyspark.sqlContext.createDataFrame` to convert the parsed lines to > a PySpark DataFrame. However, all entries are parsed as string type > regardless of what the correct data type is. > An example of my code and output can be found at: > https://gist.github.com/karenyyng/a1264d6344c54df4fcc5 -- This message was sent by Atlassian JIRA (v6.3.4#6332) --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org