a feature transformer that merges multiple columns into a vector column

Solutions on MaxInterview for a feature transformer that merges multiple columns into a vector column by the best coders in the world

Search
showing results for - "a feature transformer that merges multiple columns into a vector column"
Lina
20 Mar 2018
1# A feature transformer that merges multiple columns into a vector column
2
3df = spark.createDataFrame([(1, 0, 3)], ["a", "b", "c"])
4vecAssembler = VectorAssembler(inputCols=[
5  "a", "b", "c"], outputCol="features")
6vecAssembler.transform(df).head().features
7# DenseVector([1.0, 0.0, 3.0])
8vecAssembler.setParams(outputCol="freqs").transform(df).head().freqs
9# DenseVector([1.0, 0.0, 3.0])
10params = {vecAssembler.inputCols: [
11  "b", "a"], vecAssembler.outputCol: "vector"}
12vecAssembler.transform(df, params).head().vector
13# DenseVector([0.0, 1.0])
14vectorAssemblerPath = temp_path + "/vector-assembler"
15vecAssembler.save(vectorAssemblerPath)
16loadedAssembler = VectorAssembler.load(vectorAssemblerPath)
17loadedAssembler.transform(df).head().freqs == vecAssembler.transform(
18  df).head().freqs
19# True
similar questions
pd merge on multiple columnspython combine two columns into matrixmerge two dataframes with common columnspandas apply output multiple columnspandas pass two columns to functionpandas merge asof directioncombine dataframes with two matching columnspython transform two columns to a list combinearray concatenate columns are equalcombine int and object columns into one
queries leading to this page
a feature transformer that merges multiple columns into a vector columna feature transformer that merges multiple columns into a vector column