In Nifi, I have the following JSON:
{
"list1":[1,2,3],
"list2":[4,5]
}
I want the output to be:
{
"Final_List":[1,2,3,4,5]
}
How should I do this?
You can add a JoltTransformJSON processor with a shift type transformation such as
[
{
"operation": "shift",
"spec": {
"*": {
"*": {
"#": "Final_List"
}
}
}
}
]
Edit for special case :
If it was the case that those arrays have some common integer elements, then the above solution would yield repeating elements within the Final_List, but the following make that result having unique elements for this case too :
[
{
"operation": "shift",
"spec": {
"*": {
"*": {
"*": "Final_List.&"
}
}
}
},
{
"operation": "shift",
"spec": {
"Final_List": {
"*": {
"$": "&2"
}
}
}
},
{
"operation": "modify-overwrite-beta",
"spec": {
"*": "=toInteger(#(1,Final_List))"
}
}
]
Related
I'm looking for flattening nested JSON file into SQL ready format.
JSON file's content:
{
"ProductLine": [
"Product 1",
"Product 2"
],
"Purchase": 364,
"Cancel": [
140,
2
]
}
My current transformation:
[
{
"operation": "shift",
"spec": {
"*": {
"*": {
"#": "[#2].&2"
}
}
}
}
]
Desired output:
[
{
"ProductLine": "Product 1",
"Purchase": 364,
"Cancel": 140
},
{
"ProductLine": "Product 2",
"Cancel": 2
}
]
The difficulty is that arrays can change, sometimes "Cancel" can be an array or sometimes "Purchase" block can be nested.
You can use this spec:
If Purchase or cancel be an array or not, this works
[
{
"operation": "cardinality",
"spec": {
"*": "MANY"
}
},
{
"operation": "shift",
"spec": {
"ProductLine": {
"*": {
"*": {
"#1": "[&2].&3",
"#(3,Purchase[&1])": "[&2].Purchase",
"#(3,Cancel[&1])": "[&2].Cancel"
}
}
}
}
}
]
First, change all values to the array. Now you can loop on the ProductLine and get other fields from Purchase and Cancel.
Update: The following answer has been obtained in collaboration with Barbaros Özhan. Special thanks.
[
{
"operation": "cardinality",
"spec": {
"*": "MANY"
}
},
{
"operation": "shift",
"spec": {
"*": {
"*": {
"#": "[#2].&2"
}
}
}
}
]
We can pick Purchase at a different(outer) level such as
[
{
"operation": "shift",
"spec": {
"*": {
"*": {
"#": "[#2].&2"
}
},
"Purchase": "[#].&"// at two level less than the inner object
}
}
]
the demo one the site http://jolt-demo.appspot.com/ is
Edit : Considering array indeterminance for the attributes, you can use the following spec alternatively
[
{ //reform two separate objects
"operation": "shift",
"spec": {
"#": "orj",
"*": "non_array.&.#0[]"
}
},
{ // in order to keep the non-array values as the first component of the newly formed array(s)
"operation": "sort"
},
{
"operation": "shift",
"spec": {
"*": { //the topmost level
"*": { //level for the keys
"*": "&1[]" //match keys and values to convert non-arrays to arrays
}
}
}
},
{// pick the first component for the non-array(s)
"operation": "modify-overwrite-beta",
"spec": {
"*": {
"*": "=firstElement"
}
}
},
{ // apply the original spec after having got individual array values
"operation": "shift",
"spec": {
"*": {
"*": {
"#": "[#2].&2"
}
}
}
},
{ //get rid of the attributes with null values
"operation": "modify-overwrite-beta",
"spec": {
"*": "=recursivelySquashNulls"
}
}
]
or another straightforward alternative would be using your original spec after applying cardinality spec such as
[
{
"operation": "cardinality",
"spec": {
"*": "MANY"
}
},
{
"operation": "shift",
"spec": {
"*": {
"*": {
"#": "[#2].&2"
}
}
}
}
]
I am getting data from the database sources in the following format with capital keys.
I want to convert keys in camel case using the JOLT transformation.
json which I may be getting does not contains fixed key value pairs. Key can vary case-to-case basis
input json
[
{
"LAST_UPDT_TS": "2018-05-21 07:52:06.0",
"RTRV_TS": "2023-02-08 06:03:03.932108",
"DOC_ID": "1-102GJ8CY",
"PARENT_ASSET_DOC_ID": null,
"ASSET_STATUS": "Inactive",
"CAGE_NUM": "SUBZ6G"
},
{
"LAST_UPDT_TS": "2020-09-09 22:28:25.0",
"RTRV_TS": "2023-02-08 06:03:03.932108",
"DOC_ID": "1-102MDPE7",
"PARENT_ASSET_DOC_ID": null,
"ASSET_STATUS": "Active",
"CAGE_NUM": "012210"
}
]
Expected output
[
{
"lastUpdtTs": "2018-05-21 07:52:06.0",
"rtrvTs": "2023-02-08 06:03:03.932108",
"docId": "1-102GJ8CY",
"ParentAssetDocId": null,
"AssetStatus": "Inactive",
"CageNum": "SUBZ6G"
},
{
"lastUpdtTs": "2020-09-09 22:28:25.0",
"retrieveTimestamp": "2023-02-08 06:03:03.932108",
"docId": "1-102MDPE7",
"ParentAssetDocId": null,
"AssetStatus": "Active",
"CageNum": "012210"
}
]
You can use the following spec
[// Preparation for the next spec for assigning default "null" for null values of which the attributes won't vanish
{
"operation": "default",
"spec": {
"*": {
"*": "null"
}
}
},
{// Exchange key-value pairs
"operation": "shift",
"spec": {
"*": {
"*": {
"$": "&2.&1.#(0)"
}
}
}
},
{// Split values by underscores in order to convert them to independent arrays for each attribute
"operation": "modify-overwrite-beta",
"spec": {
"*": {
"*": {
"*": "=split('_',#(1,&))"
}
}
}
},
{// Convert all values to lowercase letters
"operation": "modify-overwrite-beta",
"spec": {
"*": {
"*": {
"*": {
"*": "=toLower"
}
}
}
}
},
{// Enumerate each components of the arrays
"operation": "shift",
"spec": {
"*": {
"*": {
"*": {
"*": "&3.&2.&1.&"
}
}
}
}
},
{// Split those components letterwise
"operation": "modify-overwrite-beta",
"spec": {
"*": {
"*": {
"*": {
"*": "=split('',#(1,&))"
}
}
}
}
},
{// Enumerate each components of the arrays
"operation": "shift",
"spec": {
"*": {
"*": {
"*": {
"*": {
"*": "&4.&3.&2.&1.&"
}
}
}
}
}
},
{
"operation": "modify-overwrite-beta",
"spec": {
"*": {// Convert only the first letters of the derived pieces to uppercase letters by use of 0th(zeroth) index
"*": {
"*": {
"*": {
"0": "=toUpper"
}
}
}
}
}
},
{// Rearrange the elements to prepare to combine the pieces
"operation": "shift",
"spec": {
"*": {
"*": {
"*": {
"*": {
"*": "&4.&3.&2"
}
}
}
}
}
},
{// Combine the pieces in order to get new values
"operation": "modify-overwrite-beta",
"spec": {
"*": {
"*": {
"*": "=join('',#(1,&))"
}
}
}
},
{
"operation": "shift",
"spec": {// Make values keys, while keys to values
"*": {
"*": {
"*": {
"$": "[&3].#(0)"
}
}
}
}
},
{
"operation": "shift",
"spec": {
"*": {// get back again the real null values
"*": {
"null": "[&2].&1",
"*": {
"#1": "[&3].&2"
}
}
}
}
}
]
There are two immediate handicaps :
Disappearing null values during key-value exchange. Used the first and the last specs to handle this issue.
Identical values which would prevent separation of keys after applied key-value exchange. An extra level of deepness added to specs in order to handle this.
i have JSON value as below :
{
"table": "table_name",
"op_type": "U",
"before": {
"AAAA": "1-1111",
"BBBB": "2022-08-31 03:57:01",
"CCCC": "2023-08-31 23:59:59"
},
"after": {
"AAAA": "1-1112",
"BBBB": "2022-08-31 10:10:34"
}
}
i want to do this how can i do?
{
"AAAA": "1-1112",
"BBBB": "2022-08-31 10:10:34",
"CCCC": "2023-08-31 23:59:59"
"changed_columns": "AAAA, BBBB"
}
AAAA: "If you have after.AAAA, take AAAA else before.AAAA", BBBB: "If you have after.BBBB, take BBBB else before.BBBB.
AND I want to add changed_columns field like this :
,"changed_columns": "AAAA, BBBB"
is there a way to do this?
You can use shift operation and getting after values for the first. and then you can using before values. So if keys match with together you have an array with two element.
Now You can get first element to getting after values with modify-overwrite-beta operations and =firstElement function.
[
{
"operation": "shift",
"spec": {
"after": {
"*": {
"$": "changed_columns[]",
"#(1,&)": "&1"
}
},
"before": {
"*": "&"
}
}
},
{
"operation": "modify-overwrite-beta",
"spec": {
"*": "=firstElement(#(1,&))",
"changed_columns": "=join(', ',#(1,changed_columns))"
}
}
]
You can use
cardinality spec lately after using "after|before" as the key in this order to determine the precedence
exchange key-value pairs consecutively twice to determine whether
really changed the components in order to form "changed_columns"
such as
[
{
// multiplex the attributes in order to generate three independent groups
"operation": "shift",
"spec": {
"after|before": { // this order is important to determine the precedence in the upcoming cardinality spec
"*": {
"#": "&",
"#(0)": "l.&",
"*": {
"#1": "f.&2"
}
}
}
}
},
{
// determine whether before vs. after values equal through this and next two specs
"operation": "modify-overwrite-beta",
"spec": {
"l": {
"*": "=lastElement(#(1,&))"
},
"f": {
"*": "=firstElement(#(1,&))"
}
}
},
{
"operation": "shift",
"spec": {
"*": "&",
"l|f": {
"*": {
"$": "lf.#(0)"
}
}
}
},
{
"operation": "shift",
"spec": {
"*": "&",
"lf": {
"*": {
"$": "&2.#(0)"
}
}
}
},
{
// construct an array from those newly formed keys
"operation": "shift",
"spec": {
"*": "&",
"lf": {
"*": {
"$": "changed_columns"
}
}
}
},
{
"operation": "modify-overwrite-beta",
"spec": {
"changed_columns": "=join(', ',#(1,&))"
}
},
{
"operation": "cardinality",
"spec": {
"*": "ONE"
}
},
{
"operation": "sort"
}
]
i have JSON value as below :
{
"table": "table_name",
"op_type": "U",
"before": {
"AAAA": "1-1111",
"BBBB": "2022-08-31 03:57:01"
},
"after": {
"AAAA": "1-1111",
"BBBB": "2022-08-31 10:10:34",
"DDDD": "2023-08-31 23:59:59"
}
}
I want to add column_names field like this :
,"changed_columns": "AAAA,BBBB,DDDD"
is there a way to do this?
You can use the following specs in which the main idea is to arrange the attributes so as to generate an array with unique elements within the an array by using successive shift transformation, then combine them within a modify transformation such as
[
{
// combine common key names for each respective values for the attributes
"operation": "shift",
"spec": {
"before|after": {
"*": {
"$": "&"
}
}
}
},
{
// construct an array from those newly formed keys
"operation": "shift",
"spec": {
"*": {
"$": "changed_columns"
}
}
},
{
// make them comma-separated
"operation": "modify-overwrite-beta",
"spec": {
"*": "=join(',',#(1,&))"
}
}
]
the demo on the site http://jolt-demo.appspot.com/ is
Edit : If your aim is to keep newly generated attribute along with the existing ones, then you can prefer using the following spec
[
{
"operation": "shift",
"spec": {
"*": "&", //else case
"before|after": {
"*": {
"$": "cc.&",
"#": "&2.&"
}
}
}
},
{
"operation": "shift",
"spec": {
"cc": {
"*": {
"$": "changed_columns"
}
},
"*": "&" //else case
}
},
{
"operation": "modify-overwrite-beta",
"spec": {
"changed_columns": "=join(',',#(1,&))"
}
}
]
I try to truncate strings in an array with JOLT. The character "^" should be removed from any string that contains it in the first position.
Sample Input:
{
"scores": [
"^aaaa",
"^bbbb",
"cccc",
"^dddd",
"eeee"
]
}
Expected Output:
{
"scores" : [ "aaaa", "bbbb", "cccc", "dddd", "eeee" ]
}
My Spec:
[
{
"operation": "modify-overwrite-beta",
"spec": {
"splittedScores": "=split('\\^',#(1,scores))",
"scores": "=join('',#(1,splittedScores))" // <--- Not working
}
}
]
Seems you'd need one extra shift transformation sandwiched between modify transformations to tame the subarrays such as
[
{
"operation": "modify-overwrite-beta",
"spec": {
"*": "=split('\\^',#(1,&))"
}
},
{
"operation": "shift",
"spec": {
"*": {
"*": "&"
}
}
},
{
"operation": "modify-overwrite-beta",
"spec": {
"*": "=join('',#(1,&))"
}
},
{
"operation": "shift",
"spec": {
"*": {
"#": "scores"
}
}
}
]
the demo on the site http://jolt-demo.appspot.com/ is
or only one shift transformation will be sufficient to use(a straightforward option) such as
[
{
"operation": "shift",
"spec": {
"*": {
"*": {
"^*": {
"$(0,1)": "&3"
},
"*": {
"$": "&3"
}
}
}
}
}
]
splitting by "$(0,1)" wildcard for the object under "^*" node. &3 represents going three levels up to grab the name of the key scores .
the demo on the site http://jolt-demo.appspot.com/ is
If ^ is being removed in the first position really matters, then the second method; if all ^ characters from any strings should be removed, then the first method should be used.