Fix reverse engineering for fk pk sharing same column (#35)

Author: tencnivel

Project.toml

@@ -1,7 +1,7 @@
 name = "PostgresORM"
 uuid = "748b5efa-ed57-4836-b183-a38105a77fdd"
 authors = ["Vincent Laugier <vincent.laugier@gmail.com>"]
-version = "0.5.7"
+version = "0.6.0"
 
 [deps]
 DataFrames = "a93c6f00-e57d-5684-b7b6-d8193f3e46c0"

src/Controller/coreORM.create.jl

@@ -95,6 +95,9 @@ function create_entity!(new_object::IEntity,
     column_names = util_getcolumns(properties_names, columns_selection_and_mapping)
     properties_values = collect(values(props))
 
+    # Dedup column names and values in case some properties are sharing the same columns
+    PostgresORMUtil.dedup_colnames_colvalues!(column_names, properties_values)
+
     # Loop over the properties of the object and
     #   build the appropriate list of columns
     if length(props) > 0
@@ -104,7 +107,7 @@ function create_entity!(new_object::IEntity,
     end
 
     # Add the prepared statement indexes
-    query_indexes = string.(collect(1:length(properties_names)))
+    query_indexes = string.(collect(1:length(column_names)))
     query_indexes = string.("\$",query_indexes)
     query_string *= (" VALUES ("
                      * join(query_indexes,",")

src/PostgresORM.jl

@@ -70,7 +70,7 @@ module PostgresORM
 
   # Provides functions to get information about the database structures
   module SchemaInfo
-      include("./SchemaInfo/SchemaInfo-def.jl") # This is only the definition of the
+      include("./SchemaInfo/_def.jl") # This is only the definition of the
                                                  #   module. See below for the actual
                                                  #   implementation.
   end #module SchemaInfo
@@ -116,7 +116,7 @@ module PostgresORM
   end # module Tool
 
   # Implementation of the SchemaInfo module
-  include("./SchemaInfo/SchemaInfo-imp.jl")
+  include("./SchemaInfo/_imp.jl")
 
 
   include("./exposed-functions-from-submodules.jl")

src/PostgresORMUtil/utils.jl

@@ -380,3 +380,59 @@ function remove_spaces_and_split(str::String)
     str = replace(str, " " => "")
     return split(str,',')
 end
+
+"""
+    dedup_colnames_colvalues!(column_names::Vector{String}, column_values::Vector)
+
+Inplace dedup column names and values by giving priority to non missing elements
+
+Eg.
+Applying function to the following arguments:
+column_names = ["exam_log_id", "in_date", "in_date", "out_time", "api_url", "in_time", "user_id"]
+column_values = ["215c81e9-e002-402e-8482-1382a65ef1e4", "2024-03-11", missing, "2024-03-11T08:27:12.363+01:00", "exam/save-values", "2024-03-11T08:27:09.363+01:00", "b7845f35-0169-488d-b8ce-111f0f07d695"]
+
+will result in a dropping the third element of both vectors
+
+"""
+function dedup_colnames_colvalues!(
+    column_names::Vector{String},
+    column_values::Vector
+)
+
+    # Function to check if a value is considered "missing" in your context
+    is_missing(value) = value === missing  # Replace 'nothing' with your definition of a missing value
+
+    # Create a dictionary to track the first occurrence and its index
+    seen = Dict{String, Int}()
+
+    # Iterate backwards to prefer the first non-missing value when deduplicating
+    for i in length(column_names):-1:1
+        name, value = column_names[i], column_values[i]
+        if !is_missing(value)
+            if haskey(seen, name)
+                # Update with non-missing value if found later
+                column_values[seen[name]] = value
+            else
+                seen[name] = i  # Track the first non-missing occurrence
+            end
+        elseif !haskey(seen, name)
+            # Even if it's missing, we need to track the first occurrence
+            seen[name] = i
+        end
+    end
+
+    # Filter the original arrays based on the indices stored in 'seen'
+    # This will also sort them based on their original order preserved by iteration order in 'seen'
+    filter_indices = sort(collect(values(seen)))
+
+    # Assign new values to the beginning of the original arrays
+    column_names[1:length(filter_indices)] = [column_names[i] for i in filter_indices]
+    column_values[1:length(filter_indices)] = [column_values[i] for i in filter_indices]
+
+    # Resize the original arrays to remove the extra elements
+    resize!(column_names, length(filter_indices))
+    resize!(column_values, length(filter_indices))
+
+    nothing
+
+end