我有一个非常奇怪的错误与火花数据框,这导致字符串被评估为时间戳。pyspark createdataframe:字符串解释为时间戳,模式混合列
这里是我的设置代码:
from datetime import datetime
from pyspark.sql import Row
from pyspark.sql.types import StructType, StructField, StringType, TimestampType
new_schema = StructType([StructField("item_id", StringType(), True),
StructField("date", TimestampType(), True),
StructField("description", StringType(), True)
])
df = sqlContext.createDataFrame([Row(description='description', date=datetime.utcnow(), item_id='id_string')], new_schema)
这给了我以下错误:
AttributeError Traceback (most recent call last) in() ----> 1 df = sqlContext.createDataFrame([Row(description='hey', date=datetime.utcnow(), item_id='id_string')], new_schema)
/home/florian/spark/python/pyspark/sql/context.pyc in createDataFrame(self, data, schema, samplingRatio, verifySchema) 307 Py4JJavaError: ... 308 """ --> 309 return self.sparkSession.createDataFrame(data, schema, samplingRatio, verifySchema) 310 311 @since(1.3)
/home/florian/spark/python/pyspark/sql/session.pyc in createDataFrame(self, data, schema, samplingRatio, verifySchema) 522 rdd, schema = self._createFromRDD(data.map(prepare), schema, samplingRatio) 523 else: --> 524 rdd, schema = self._createFromLocal(map(prepare, data), schema) 525 jrdd = self._jvm.SerDeUtil.toJavaArray(rdd._to_java_object_rdd()) 526 jdf = self._jsparkSession.applySchemaToPythonRDD(jrdd.rdd(), schema.json())
/home/florian/spark/python/pyspark/sql/session.pyc in _createFromLocal(self, data, schema) 397 398 # convert python objects to sql data --> 399 data = [schema.toInternal(row) for row in data] 400 return self._sc.parallelize(data), schema 401
/home/florian/spark/python/pyspark/sql/types.pyc in toInternal(self, obj) 574 return tuple(f.toInternal(obj.get(n)) for n, f in zip(self.names, self.fields)) 575 elif isinstance(obj, (tuple, list)): --> 576 return tuple(f.toInternal(v) for f, v in zip(self.fields, obj)) 577 elif hasattr(obj, "dict"): 578 d = obj.dict
/home/florian/spark/python/pyspark/sql/types.pyc in ((f, v)) 574 return tuple(f.toInternal(obj.get(n)) for n, f in zip(self.names, self.fields)) 575 elif isinstance(obj, (tuple, list)): --> 576 return tuple(f.toInternal(v) for f, v in zip(self.fields, obj)) 577 elif hasattr(obj, "dict"): 578 d = obj.dict
/home/florian/spark/python/pyspark/sql/types.pyc in toInternal(self, obj) 434 435 def toInternal(self, obj): --> 436 return self.dataType.toInternal(obj) 437 438 def fromInternal(self, obj):
/home/florian/spark/python/pyspark/sql/types.pyc in toInternal(self, dt) 188 def toInternal(self, dt): 189 if dt is not None: --> 190 seconds = (calendar.timegm(dt.utctimetuple()) if dt.tzinfo 191 else time.mktime(dt.timetuple())) 192 return int(seconds * 1e6 + dt.microsecond)
AttributeError: 'str' object has no attribute 'tzinfo'
这看起来好像一个字符串传递给TimestampType.toInternal()
的真奇怪的是,这个数据帧会产生相同的错误:
df = sqlContext.createDataFrame([Row(description='hey', date=None, item_id='id_string')], new_schema)
,而这一个工程:
df = sqlContext.createDataFrame([Row(description=None, date=datetime.now(), item_id='id_string')], new_schema)
而这一次的作品,以及:
df = sqlContext.createDataFrame([Row(description=None, date=datetime.now(), item_id=None)], new_schema)
对于我来说,这意味着现在的pyspark莫名其妙地把价值从“ITEM_ID”入列“日期“并因此产生这个错误。 我做错了什么?这是数据框架内的错误吗?
信息: 我使用pyspark 2.0.1
编辑:
df = sqlContext.createDataFrame([Row(description=None, date=datetime.now(), item_id=None)], new_schema)
df.first()
Row(item_id=u'java.util.GregorianCalendar[time=?,areFieldsSet=false,areAllFieldsSet=false,lenient=true,zone=sun.util.calendar.ZoneInfo[id="Etc/UTC",offset=0,dstSavings=0,useDaylight=false,transitions=0,lastRule=null],firstDayOfWeek=1,minimalDaysInFirstWeek=1,ERA=?,YEAR=2017,MONTH=1,WEEK_OF_YEAR=?,WEEK_OF_MONTH=?,DAY_OF_MONTH=3,DAY_OF_YEAR=?,DAY_OF_WEEK=?,DAY_OF_WEEK_IN_MONTH=?,AM_PM=1,HOUR=3,HOUR_OF_DAY=15,MINUTE=19,SECOND=30,MILLISECOND=85,ZONE_OFFSET=?,DST_OFFSET=?]', date=None, description=None)