[英]Unable to remove “\” from the json nested attributes using spark scala
我正在研究一個FHIR資源，在該資源中我得到一個 JSON 數據，如下所示：
{
"appointmentRef": "Appointment/12213#4200",
"encounterLengh": "2",
"billingAccount": "savingsAccount",
"hospitalization": "{\"preAdmissionIdentifierSystem\":\"https://system123445.html\",\"preAdmissionIdentifierValue\":\"pqr\",\"origin\":\"hospital\",\"admitSourceCode\":\"outp\",\"admitSourceReason\":\"some thing\",\"eid\":200,\"destination\":\"hospital\"}",
"resourceType": "Encounter",
"priority": "abc",
"status": "triaged",
"eid": "200",
"subject": "Patient/435"
}
因此，以前對於位於根級別的屬性，例如appointmentRef參考等。他們在 R.HS 上也有"\" ，我可以通過我的代碼將其刪除。 但是，從上面的數據可以看出，對於嵌套屬性，我的代碼不起作用。
rowList.groupBy(row => row.key).foreach(rowList => {
import com.google.gson.{Gson, JsonObject}
val map: Map[String, String] = mutable.Map()
rowList._2.foreach(row => {
LOGGER.debug(s"row == $row")
if (Utility.isBlank(row.jsonElementTag)) {
val convertedObject = new Gson().fromJson(row.value, classOf[JsonObject])
val itr = convertedObject.entrySet().iterator()
while (itr.hasNext) {
val next = itr.next()
val value = next.getValue.getAsString
val key = next.getKey
LOGGER.debug(s"key-- $key value --$value")
map.put(key, value)
}
}
else {
val convertedObject = new Gson().fromJson(row.value, classOf[JsonObject])
LOGGER.debug(s"convertedObject == $convertedObject")
if (null != map.get(row.jsonElementTag).getOrElse(null)) {
LOGGER.debug("map.get(row.jsonElementTag).get === "+row.jsonElementTag +" "+map.get(row.jsonElementTag).get)
var array: JsonArray = new JsonArray
val mapElement = new Gson().fromJson(map.get(row.jsonElementTag).get, classOf[JsonObject])
array.add(mapElement)
array.add(convertedObject)
map.put(row.jsonElementTag, array.toString)
}
else {
map.put(row.jsonElementTag, convertedObject.toString)
}
}
})
我只是從數據框中獲取行並遍歷行，將其作為字符串，並將其放入鍵值對中。 if循環將為父級屬性運行， else-if循環將為嵌套屬性執行。
我什至嘗試了更簡單的replace("\","")方法，但沒有奏效。 那么，如何從嵌套屬性中刪除"\" ？
我expected的 output 在我的嵌套 JSON 屬性中應該沒有"\" 。
hospitalization列是字符串類型，它包含json object 。 要將字符串提取或轉換為 json，請根據該列中的數據准備schema 。
檢查下面的代碼。
scala> import org.apache.spark.sql.types._
import org.apache.spark.sql.types._
scala> val schema = DataType.fromJson("""{"type":"struct","fields":[{"name":"admitSourceCode","type":"string","nullable":true,"metadata":{}},{"name":"admitSourceReason","type":"string","nullable"
:true,"metadata":{}},{"name":"destination","type":"string","nullable":true,"metadata":{}},{"name":"eid","type":"long","nullable":true,"metadata":{}},{"name":"origin","type":"string","nullable":tr
ue,"metadata":{}},{"name":"preAdmissionIdentifierSystem","type":"string","nullable":true,"metadata":{}},{"name":"preAdmissionIdentifierValue","type":"string","nullable":true,"metadata":{}}]}""").
asInstanceOf[StructType]
scala> df.withColumn("hospitalization",from_json($"hospitalization",schema)).printSchema
root
|-- appointmentRef: string (nullable = true)
|-- billingAccount: string (nullable = true)
|-- eid: string (nullable = true)
|-- encounterLengh: string (nullable = true)
|-- hospitalization: struct (nullable = true)
| |-- admitSourceCode: string (nullable = true)
| |-- admitSourceReason: string (nullable = true)
| |-- destination: string (nullable = true)
| |-- eid: long (nullable = true)
| |-- origin: string (nullable = true)
| |-- preAdmissionIdentifierSystem: string (nullable = true)
| |-- preAdmissionIdentifierValue: string (nullable = true)
|-- priority: string (nullable = true)
|-- resourceType: string (nullable = true)
|-- status: string (nullable = true)
|-- subject: string (nullable = true)
scala> df.withColumn("hospitalization",from_json($"hospitalization",schema)).show(false)
+----------------------+--------------+---+--------------+---------------------------------------------------------------------------+--------+------------+-------+-----------+
|appointmentRef |billingAccount|eid|encounterLengh|hospitalization |priority|resourceType|status |subject |
+----------------------+--------------+---+--------------+---------------------------------------------------------------------------+--------+------------+-------+-----------+
|Appointment/12213#4200|savingsAccount|200|2 |[outp, some thing, hospital, 200, hospital, https://system123445.html, pqr]|abc |Encounter |triaged|Patient/435|
+----------------------+--------------+---+--------------+---------------------------------------------------------------------------+--------+------------+-------+-----------+
更新
創建了小助手 class 以在沒有架構的情況下提取或轉換 json。
import org.apache.spark.sql.functions._
import org.apache.spark.sql.expressions._
import org.json4s.JsonDSL._
import org.json4s._
import org.json4s.jackson.JsonMethods._
val append = udf((rowId: Long,json: String) => {
compact(render(Map("rowId" -> parse(rowId.toString),"data" ->parse(json))))
})
implicit class DFHelper(df: DataFrame) {
import df.sparkSession.implicits._
def parseJson = df.sparkSession.read.option("multiLine","true").json(df.map(_.getString(0)))
//Convert string to json object or array of json object
def extract(column: Column) = {
val updatedDF = df.withColumn("rowId",row_number().over(Window.orderBy(lit(1))))
val parsedDF = updatedDF.filter(column.isNotNull)
.select(append($"rowid",column).as("row"))
.parseJson
updatedDF.join(
parsedDF.select($"rowId",$"data".as(column.toString())),
updatedDF("rowId") === parsedDF("rowId"),
"left"
)
.drop("rowId") // Deleting added rowId column.
}
}
scala> df.extract($"hospitalization").printSchema()
root
|-- appointmentRef: string (nullable = true)
|-- billingAccount: string (nullable = true)
|-- eid: string (nullable = true)
|-- encounterLengh: string (nullable = true)
|-- hospitalization: string (nullable = true)
|-- priority: string (nullable = true)
|-- resourceType: string (nullable = true)
|-- status: string (nullable = true)
|-- subject: string (nullable = true)
|-- hospitalization: struct (nullable = true)
| |-- admitSourceCode: string (nullable = true)
| |-- admitSourceReason: string (nullable = true)
| |-- destination: string (nullable = true)
| |-- eid: long (nullable = true)
| |-- encounterLengh: string (nullable = true)
| |-- origin: string (nullable = true)
| |-- preAdmissionIdentifierSystem: string (nullable = true)
| |-- preAdmissionIdentifierValue: string (nullable = true)
scala> df.extract($"hospitalization").show(false)
+----------------------+--------------+---+--------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+--------+------------+-------+-----------+------------------------------------------------------------------------------+
|appointmentRef |billingAccount|eid|encounterLengh|hospitalization |priority|resourceType|status |subject |hospitalization |
+----------------------+--------------+---+--------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+--------+------------+-------+-----------+------------------------------------------------------------------------------+
|Appointment/12213#4200|savingsAccount|200|1 |{"encounterLengh": "1","preAdmissionIdentifierSystem":"https://system123445.html","preAdmissionIdentifierValue":"pqr","origin":"hospital","admitSourceCode":"outp","admitSourceReason":"some thing","eid":200,"destination":"hospital"}|abc |Encounter |triaged|Patient/435|[outp, some thing, hospital, 200, 1, hospital, https://system123445.html, pqr]|
+----------------------+--------------+---+--------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+--------+------------+-------+-----------+------------------------------------------------------------------------------+
也許這有幫助-
val data =
"""
|{
| "appointmentRef": "Appointment/12213#4200",
| "encounterLengh": "2",
| "billingAccount": "savingsAccount",
| "hospitalization": "{\"preAdmissionIdentifierSystem\":\"https://system123445.html\",\"preAdmissionIdentifierValue\":\"pqr\",\"origin\":\"hospital\",\"admitSourceCode\":\"outp\",\"admitSourceReason\":\"some thing\",\"eid\":200,\"destination\":\"hospital\"}",
| "resourceType": "Encounter",
| "priority": "abc",
| "status": "triaged",
| "eid": "200",
| "subject": "Patient/435"
|}
""".stripMargin
val ds = Seq(data).toDF()
ds.show(false)
ds.printSchema()
/**
* +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
* |value |
* +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
* |
* {
* "appointmentRef": "Appointment/12213#4200",
* "encounterLengh": "2",
* "billingAccount": "savingsAccount",
* "hospitalization": "{\"preAdmissionIdentifierSystem\":\"https://system123445.html\",\"preAdmissionIdentifierValue\":\"pqr\",\"origin\":\"hospital\",\"admitSourceCode\":\"outp\",\"admitSourceReason\":\"some thing\",\"eid\":200,\"destination\":\"hospital\"}",
* "resourceType": "Encounter",
* "priority": "abc",
* "status": "triaged",
* "eid": "200",
* "subject": "Patient/435"
* }
* |
* +------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
*
* root
* |-- value: string (nullable = true)
*/
\替換為'' （空字符串） ds.withColumn("value", translate($"value", "\\", ""))
.show(false)
/**
* +----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
* |value |
* +----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
* |
* {
* "appointmentRef": "Appointment/12213#4200",
* "encounterLengh": "2",
* "billingAccount": "savingsAccount",
* "hospitalization": "{"preAdmissionIdentifierSystem":"https://system123445.html","preAdmissionIdentifierValue":"pqr","origin":"hospital","admitSourceCode":"outp","admitSourceReason":"some thing","eid":200,"destination":"hospital"}",
* "resourceType": "Encounter",
* "priority": "abc",
* "status": "triaged",
* "eid": "200",
* "subject": "Patient/435"
* }
* |
* +----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
*/
無法使用 Spark/Scala 從 JSON 嵌套鍵值對創建列和值
[英]Unable to create the column and value from the JSON nested key Value Pair using Spark/Scala
如何動態地使用 spark scala 將 spark dataframe 轉換為嵌套 json
[英]How to Convert spark dataframe to nested json using spark scala dynamically
使用Spark Scala將嵌套JSON中的字符串變量轉換為datetime
[英]Convert a string variable in nested JSON to datetime using Spark Scala
展平任何嵌套的 json 字符串並使用 spark scala 轉換為數據幀
[英]Flatten any nested json string and convert to dataframe using spark scala
聲明:本站的技術帖子網頁，遵循CC BY-SA 4.0協議，如果您需要轉載，請注明本站網址或者原文地址。任何問題請咨詢:yoyou2525@163.com.