在AWS Glue中更改动态框架的列数据类型需要进行以下步骤：

1. 创建动态框架
2. 通过select_fields（选择特定列）和resolveChoice（从dynamicframe为Dataframe转换并指定列数据类型）方法选定特定列和更改数据类型如下：

# create a dynamic frame
dynamic_frame = glueContext.create_dynamic_frame.from_catalog(database='mydb', table_name='mytable')

# select specific column
df = dynamic_frame.select_fields(['col1'])

# resolve choice as per the data type you need
data_frame = df.toDF().withColumn('col1', df['col1'].cast('integer'))

# convert back to dynamic frame
dyf = DynamicFrame.fromDF(data_frame, glueContext, 'dyf')

3. 在更改后，将数据写入指定数据源

glueContext.write_dynamic_frame.from_jdbc_conf(
    frame = dyf, 
    catalog_connection = "myRedshiftConnection", 
    connection_options = {"dbtable": "mytable", "database": "mydb"}, 
    redshift_tmp_dir = args["TempDir"], 
    transformation_ctx = "datasink")