ValueError: длина объекта (1) не соответствует длине полей (2)

Я запускаю скрипт Python на Spark, но получаю следующую ошибку: ValueError: длина объекта (1) не соответствует длине полей (2). Я перечислю код, который я запускаю, и скриншот ошибки:

conf = SparkConf().setMaster("local").setAppName("ApplySentimentModel")
sc = SparkContext(conf = conf)
spark = SparkSession(sc)

# read the sentiment tsv file into an rdd and split it based upon tab
# the test file should be in the following format
# label->tab->tweet
lines = sc.textFile("test_file.tsv").map(lambda x: x.split("\t"))
# define the schema
schema = StructType([StructField("target", StringType(), True), StructField("tweet", StringType(), True)])
# create dataframe from rdd
docs = spark.createDataFrame(lines, schema)

# define the processing pipeline (tokenize->tf->idf->label_indexer)
tokenizer = Tokenizer(inputCol = "tweet", outputCol = "words")
hashtf = HashingTF(inputCol = "words", outputCol='tf')
idf = IDF(inputCol='tf', outputCol = "features", minDocFreq=5)
indexedLabel = StringIndexer(inputCol = "target", outputCol = "label")
pipeline = Pipeline(stages=[tokenizer, hashtf, idf, indexedLabel])

# apply the pipeline to the testing datasets
pipelineFit = pipeline.fit(docs)
test_df = pipelineFit.transform(docs)

# load the trained model
lrModel = LogisticRegressionModel.load('lrModel.model')
# apply the model to predict the sentiment
predictions = lrModel.transform(test_df)

# compute the accuracy of predictions
accuracy = predictions.filter(predictions.label == predictions.prediction).count() / float(docs.count())
print("****************************************\n")
print("Prediction accuracy " + str(accuracy) + "\n")
print("****************************************\n")

# save the testing data, and the the target and predicted labels (0 -> positive, 1 -> negative)
predictions.select("tweet", "label", "prediction").write.format('csv').option('header',True).mode('overwrite').option('sep',',').save('predictions')

sc.stop()