Skip to content
This repository has been archived by the owner on Sep 25, 2024. It is now read-only.

Latest commit

 

History

History
287 lines (229 loc) · 10.1 KB

config-map-merging.md

File metadata and controls

287 lines (229 loc) · 10.1 KB
Raw

Configuration map merging

The jenkins-pipeline-library contains the helper MapUtils.groovy which is heavily used to integrate custom changes into the configuration map.

MapUtils.groovy will merge 2-n Maps recursively together and returns the merged result without modifying the original maps.

Table of contents

Merge rules

The MapUtils tool merges maps using the following rules.

Add

Keys that are unique in each map to merge are just "merged" together.

map1 = ["property_map1": "value_map1"]
map2 = ["property_map2": "value_map2"]
result1 = MapUtils.merge(map1, map2)

/*
result1 = [
  "property_map1" : "value_map1",
  "property_map2" : "value_map2"
]
 */
  • property_map2 did not exist in map1 so it was just added.

Overwrite

When items in the map can not be merged they are overwritten by the map "to the right".

map1 = ["int": 1, "string": "1", "list": [1], "map": ["key": "map1Value"]]
map2 = ["int": 2, "string": "2", "list": 2, "map": 2]
result1 = MapUtils.merge(map1, map2)
result2 = MapUtils.merge(map2, map1)

/*
result1 = [ "int" : 2, "string" : "2", "list" : 2, "map" : 2 ]
result2 = [ "int" : 1, "string" : "1", "list" : [1], "map" : [ "key" : "map1Value" ] ]
*/
  • result1: When map2 was merged into map1 all keys were present in both maps, but the datatypes were not mergeable so the items from map2 replaced the items from map1
  • result2: This is the opposite to result1 but here map1 was merged into map2 so the values from map1 replaced the items from map2

Merge

The following datatypes support merging:

  • java.util.List
  • java.util.Map

So when the datatype is one of the above ones and the datatype is equal for an item key the item is merged.

Merging Lists

map1 = ["list": [1, 0]]
map2 = ["list": [2, 0]]
result1 = MapUtils.merge(map1, map2)
result2 = MapUtils.merge(map2, map1)

/*
result1 = [ "list" : [ 1, 0, 2 ] ]
result2 = [ "list" : [ 2, 0, 1 ] ]
*/

As you can see both results contain the same elements in different order, depending if you merge map2 into map1 or the other way around.

Merging Maps

Maps will be merged recursively following the rules from above for Adding, Overwriting and Merging items.

map1 = [lvl1: [map1Item: "map1Value", list: [1], int: 1, str: "1", "lvl2": [map1: "1"]]]
map2 = [lvl1: [map2Item: "map2Value", list: [2], int: 2, str: "2", "lvl2": [map2: "2"]]]
result1 = MapUtils.merge(map1, map2)
result2 = MapUtils.merge(map2, map1)
/*
result1 = [lvl1:[map1Item: "map1Value", list: [1, 2], int: 2, str: "2", lvl2 : [map1: "1", map2: "2"], map2Item: "map2Value" ]]
result2 = [lvl1:[map2Item: "map2Value", list: [2, 1], int: 1, str: "1", lvl2 : [map2: "2", map1: "1"], map1Item: "map1Value" ]]
 */
result1 explanation
  • The lvl1 item from map2 was integrated into the lvl1 item from map1
  • The list from map1 and map2 was merged because the items from both maps were of the type java.util.List
  • The items int and str from map2 replaced the corresponding items from map1
  • The map item lvl2 from map2 merged with the corresponding item from map1
  • The item map2Item was added
result2 explanation

The result2 is like the result1 but items from map1 had precedence before items from map2.

Merge order

The passed maps will be processed from the left to the right which means that the second Map will overwrite the first Map.

Map map1 = [ "source" : "map1", "str1" : "hello map1" ] 
Map map2 = [ "source" : "map2", "str2" : "hello map2" ] 
Map map3 = [ "source" : "map3", "str3" : "hello map3" ] 

Map result1 = MapUtils.merge(map1, map2) 
Map result2 = MapUtils.merge(map1, map2, map3) 
/*
result1 = [ 
            "source" : "map1", 
            "map1" : "hello map1",
            "map2" : "hello map2" 
          ]
result2 = [ 
            "source" : "map1", 
            "map1" : "hello map1",
            "map2" : "hello map2",
            "map2" : "hello map3" 
          ]
*/

For result1 the merge was applied as follows

  • map1 and map2 both contained source and source is a simple data type (string) so the value from map2 replaced the value of map1
  • str2 from map2 was simply added to the result because is did not exist in map1

For result2 the merge was applied as follows

  • map1 and map2 both contained source and source is a simple data type (string) so the value from map2 replaced the value of map1
  • str2 from map2 was simply added to the result because is did not exist in map1
  • map3 also contained source and source is a simple data type (string) so the value from the merge of map1 and map2 were replaced by the value of map3
  • str3 from map3 was simply added to the result because is did not exist in map1 nor map2

Merge Modes

With version 1.6.0 of the jenkins pipeline library merge modes were introduced to provide more control about the merge behavior.

The existing, default merge mode is MapMergeMode.MERGE. The corresponding config constant is ConfigConstants.MAP_MERGE_MODE.

The used map merge mode is retrieved in the following order:

  • use the MapMergeMode from the existing / "left" map, if defined
  • use the MapMergeMode from the incoming / "right" map, if defined
  • use the MapMergeMode.MERGE as default

Example definition:

import io.wcm.devops.jenkins.pipeline.utils.ConfigConstants
import io.wcm.devops.jenkins.pipeline.utils.maps.MapMergeMode 

Map config = [
    (ConfigConstants.MAP_MERGE_MODE) : MapMergeMode.MERGE
  ]

MapMergeMode.MERGE

This is the default merge mode and follows the merge rules.

MapMergeMode.SKIP

SKIP the merge, keep the existing value, do not overwrite items in the "left" map with items in the "right" map.

map1 = [int: 1, string: "1", list: [1], map: [map1Key: "map1Value"], map1Item: "map1Item", (MAP_MERGE_MODE): MapMergeMode.SKIP]
map2 = [int: 2, string: "2", list: [2], map: [map2Key: "map2Value"], map2Item: "map2Item" ]
result1 = MapUtils.merge(map1, map2)
result2 = MapUtils.merge(map2, map1)
/*
result1 = [int: 1, string: "1", list: [1], map: [map1Key: "map1Value"], map1Item: "map1Item", mapMergeMode: "SKIP", map2Item: "map2Item"]
result2 = [int: 2, string: "2", list: [2], map: [map2Key: "map2Value"], map2Item: "map2Item", map1Item: "map1Item", mapMergeMode: "SKIP"]
 */

For result1 the merge was applied as follows:

  • each value for a key present in map1 and map2 was taken from map1 - so the map and the list item were not merged!
  • All other "non-overlapping" items were added (e.g. map2Item)

For result2 the merge was applied as follows:

  • each value for a key present in map1 and map2 was taken from map2 - so the map and the list item were not merged!
  • All other "non-overlapping" items were added (e.g. map1Item and mapMergeMode)

MapMergeMode.REPLACE

This mode will replace all existing / overlapping keys in the map to the left with the items from the map to the right. So this mode is the opposite of MapMergeMode.SKIP.

map1 = [int: 1, string: "1", list: [1], map: [map1Key: "map1Value"], map1Item: "map1Item", (MAP_MERGE_MODE): MapMergeMode.REPLACE]
map2 = [int: 2, string: "2", list: [2], map: [map2Key: "map2Value"], map2Item: "map2Item"]
result1 = MapUtils.merge(map1, map2)
result2 = MapUtils.merge(map2, map1)
/*
result1 = [int: 2, string: "2", list: [2], map: [ map2Key: "map2Value"], map1Item: "map1Item", mapMergeMode: "REPLACE", map2Item: "map2Item"]
result2 = [int: 1, string: "1", list: [1], map: [ map1Key: "map1Value"], map2Item: "map2Item", map1Item: "map1Item" , mapMergeMode: "REPLACE"]
*/

For result1 the merge was applied as follows:

  • each value for a key present in map1 and map2 was taken from map2 - so the map and the list item were not merged!
  • All other "non-overlapping" items were added (e.g. map2Item)

For result2 the merge was applied as follows:

  • each value for a key present in map1 and map2 was taken from map1 - so the map and the list item were not merged!
  • All other "non-overlapping" items were added (e.g. map1Item and mapMergeMode)

MapMergeMode scope

The MapMergeMode can only be defined in items of the type java.util.Map and they are only valid for the current map item / level.

map1 = [merge: [key_merge: ["map1"]], skip: [key_skip: ["map1"], (MAP_MERGE_MODE): MapMergeMode.SKIP], replace: [key_replace: ["map1"], (MAP_MERGE_MODE): MapMergeMode.REPLACE]]
map2 = [merge: [key_merge: ["map2"]], skip: [key_skip: ["map2"]], replace: [key_replace: ["map2"]]]
result1 = MapUtils.merge(map1, map2)
result2 = MapUtils.merge(map2, map1)
/*
result1 = [merge: [key_merge: ["map1", "map2"]], skip: [key_skip:["map1"], mapMergeMode:SKIP], replace: [key_replace: ["map2"], mapMergeMode: "REPLACE"]]
result2 = [merge: [key_merge: ["map2", "map1"]], skip: [key_skip:["map2"], mapMergeMode:SKIP], replace: [key_replace: ["map1"], mapMergeMode: "REPLACE"]]
*/
  • The merge item used the default merge mode (MapMergeMode.MERGE)
  • The skip item used the merge mode (MapMergeMode.SKIP)
    • In result1 the values from the incoming map map2 were skipped so the values from map1 were retained
    • In result2 the values from the incoming map map1 were skipped so the values from map2 were retained
  • The replace item used the merge mode (MapMergeMode.REPLACE)
    • In result1 the values from the incoming map map2 were used so the values from map1 were replaced
    • In result2 the values from the incoming map map1 were used so the values from map2 were replaced