diff --git a/src/create-model.jl b/src/create-model.jl
index bc69a338..fe30cfa0 100644
--- a/src/create-model.jl
+++ b/src/create-model.jl
@@ -17,6 +17,7 @@ function create_model!(energy_problem; kwargs...)
         dataframes = energy_problem.dataframes
         sets = create_sets(graph, years)
         energy_problem.model = @timeit to "create_model" create_model(
+            energy_problem.db_connection,
             graph,
             sets,
             variables,
@@ -44,6 +45,7 @@ end
 Create the energy model given the `graph`, `representative_periods`, dictionary of `dataframes` (created by [`construct_dataframes`](@ref)), timeframe, and groups.
 """
 function create_model(
+    connection,
     graph,
     sets,
     variables,
@@ -115,6 +117,7 @@ function create_model(
         incoming_flow_storage_inter_rp_balance,
         outgoing_flow_storage_inter_rp_balance,
     ) = add_expressions_to_dataframe!(
+        connection,
         dataframes,
         variables,
         model,
diff --git a/src/model-preparation.jl b/src/model-preparation.jl
index 960cfd7e..7777e4c7 100644
--- a/src/model-preparation.jl
+++ b/src/model-preparation.jl
@@ -62,8 +62,14 @@ function construct_dataframes(
 
     tmp_create_constraints_indices(connection)
 
+    # WIP: Can these queries be left undordered by the end of the refactor?
     # WIP: highest_in_out is not included in constraints_partition anymore
-    dataframes[:highest_in_out] = TulipaIO.get_table(connection, "cons_indices_highest_in_out")
+    dataframes[:highest_in_out] =
+        DuckDB.query(
+            connection,
+            "SELECT * FROM cons_indices_highest_in_out
+            ORDER BY asset, year, rep_period, time_block_start",
+        ) |> DataFrame
     dataframes[:highest_in_out].timesteps_block = map(
         r -> r[1]:r[2],
         zip(
@@ -75,7 +81,12 @@ function construct_dataframes(
 
     # WIP: highest_in is not included in constraints_partition anymore
     tmp_create_constraints_indices(connection)
-    dataframes[:highest_in] = TulipaIO.get_table(connection, "cons_indices_highest_in")
+    dataframes[:highest_in] =
+        DuckDB.query(
+            connection,
+            "SELECT * FROM cons_indices_highest_in
+            ORDER BY asset, year, rep_period, time_block_start",
+        ) |> DataFrame
     dataframes[:highest_in].timesteps_block = map(
         r -> r[1]:r[2],
         zip(dataframes[:highest_in].time_block_start, dataframes[:highest_in].time_block_end),
@@ -480,6 +491,7 @@ function add_expression_terms_inter_rp_constraints!(
 end
 
 function add_expressions_to_dataframe!(
+    connection,
     dataframes,
     variables,
     model,
@@ -512,15 +524,38 @@ function add_expressions_to_dataframe!(
             use_highest_resolution = false,
             multiply_by_duration = true,
         )
-        add_expression_terms_intra_rp_constraints!(
-            dataframes[:highest_in_out],
-            dataframes[:flows],
-            expression_workspace,
-            representative_periods,
-            graph;
-            use_highest_resolution = true,
-            multiply_by_duration = false,
-        )
+        # WIP: Changing the function below
+        # add_expression_terms_intra_rp_constraints!(
+        #     dataframes[:highest_in_out],
+        #     dataframes[:flows],
+        #     expression_workspace,
+        #     representative_periods,
+        #     graph;
+        #     use_highest_resolution = true,
+        #     multiply_by_duration = false,
+        # )
+        dataframes[:highest_in_out].incoming_flow = JuMP.AffExpr[
+            if length(row.indices) > 0
+                sum(unique(variables[:flow].container[row.indices]))
+            else
+                JuMP.AffExpr(0.0)
+            end for row in DuckDB.query(
+                connection,
+                "SELECT * FROM highest_in_out_incoming
+                ORDER BY asset, year, rep_period, time_block_start",
+            )
+        ]
+        dataframes[:highest_in_out].outgoing_flow = JuMP.AffExpr[
+            if length(row.indices) > 0
+                sum(unique(variables[:flow].container[row.indices]))
+            else
+                JuMP.AffExpr(0.0)
+            end for row in DuckDB.query(
+                connection,
+                "SELECT * FROM highest_in_out_outgoing
+                ORDER BY asset, year, rep_period, time_block_start",
+            )
+        ]
         add_expression_terms_intra_rp_constraints!(
             dataframes[:highest_in],
             dataframes[:flows],
diff --git a/src/tmp.jl b/src/tmp.jl
index 6765528d..e83030c5 100644
--- a/src/tmp.jl
+++ b/src/tmp.jl
@@ -124,7 +124,7 @@ function tmp_create_partition_tables(connection)
 
     DBInterface.execute(
         connection,
-        "CREATE OR REPLACE TABLE flow_time_resolution(
+        "CREATE OR REPLACE TEMP TABLE t_flow_time_resolution(
             from_asset STRING,
             to_asset STRING,
             year INT,
@@ -135,7 +135,7 @@ function tmp_create_partition_tables(connection)
         )",
     )
 
-    appender = DuckDB.Appender(connection, "flow_time_resolution")
+    appender = DuckDB.Appender(connection, "t_flow_time_resolution")
     for row in TulipaIO.get_table(Val(:raw), connection, "explicit_flows_rep_periods_partitions")
         durations = if row.specification == "uniform"
             step = parse(Int, row.partition)
@@ -157,6 +157,17 @@ function tmp_create_partition_tables(connection)
         _append_given_durations(appender, row, durations)
     end
     DuckDB.close(appender)
+
+    DBInterface.execute(
+        connection,
+        "CREATE OR REPLACE TABLE flow_time_resolution
+        AS
+            SELECT
+                *,
+                row_number() OVER (ORDER BY from_asset, to_asset, year, rep_period, time_block_start) AS variable_index
+            FROM t_flow_time_resolution
+        ",
+    )
 end
 
 function tmp_create_constraints_indices(connection)
@@ -265,6 +276,9 @@ function tmp_create_constraints_indices(connection)
         WHERE main.type in ('storage')
         ",
     )
+
+    # WIP: For now, the incoming and outgoing expressions are being computed here
+    tmp_create_expressions(connection)
 end
 
 """
@@ -303,3 +317,99 @@ function tmp_example_of_flow_expression_problem()
 
     return connection, ep
 end
+
+function tmp_create_expressions(connection)
+    DuckDB.execute(
+        connection,
+        "CREATE OR REPLACE TABLE t_incoming_nonzero AS
+            SELECT
+                t_flows.from_asset,
+                t_flows.to_asset,
+                t_flows.year,
+                t_flows.rep_period,
+                t_flows.variable_index,
+                t_cons.time_block_start,
+                greatest(0, 1 +
+                    least(t_flows.time_block_end,t_cons.time_block_end)
+                    - greatest(t_flows.time_block_start,t_cons.time_block_start)) AS duration
+            FROM flow_time_resolution AS t_flows
+            INNER JOIN cons_indices_highest_in_out AS t_cons
+                ON t_flows.to_asset=t_cons.asset
+                    AND t_flows.year=t_cons.year
+                    AND t_flows.rep_period=t_cons.rep_period
+            WHERE duration > 0",
+    )
+
+    DuckDB.execute(
+        connection,
+        "CREATE OR REPLACE TABLE t_outgoing_nonzero AS
+            SELECT
+                t_flows.from_asset,
+                t_flows.to_asset,
+                t_flows.year,
+                t_flows.rep_period,
+                t_flows.variable_index,
+                t_cons.time_block_start,
+                greatest(0, 1 +
+                    least(t_flows.time_block_end,t_cons.time_block_end)
+                    - greatest(t_flows.time_block_start,t_cons.time_block_start)) AS duration
+            FROM flow_time_resolution AS t_flows
+            INNER JOIN cons_indices_highest_in_out AS t_cons
+                ON t_flows.from_asset=t_cons.asset
+                    AND t_flows.year=t_cons.year
+                    AND t_flows.rep_period=t_cons.rep_period
+            WHERE duration > 0",
+    )
+
+    DuckDB.execute(
+        connection,
+        "CREATE OR REPLACE TABLE highest_in_out_incoming AS
+        SELECT
+            t1.asset, t1.year, t1.rep_period, t1.time_block_start,
+            COALESCE(indices, []) AS indices,
+            COALESCE(durations, []) AS durations,
+        FROM cons_indices_highest_in_out AS t1
+        LEFT JOIN (
+            SELECT
+                to_asset as asset,
+                year,
+                rep_period,
+                time_block_start,
+                array_agg(variable_index) AS indices,
+                array_agg(duration) AS durations,
+            FROM t_incoming_nonzero
+            GROUP BY to_asset, year, rep_period, time_block_start
+        ) AS t2
+        ON t1.asset=t2.asset
+            AND t1.year=t2.year
+            AND t1.rep_period=t2.rep_period
+            AND t1.time_block_start=t2.time_block_start
+        ",
+    )
+
+    DuckDB.execute(
+        connection,
+        "CREATE OR REPLACE TABLE highest_in_out_outgoing AS
+        SELECT
+            t1.asset, t1.year, t1.rep_period, t1.time_block_start,
+            COALESCE(indices, []) AS indices,
+            COALESCE(durations, []) AS durations,
+        FROM cons_indices_highest_in_out AS t1
+        LEFT JOIN (
+            SELECT
+                from_asset as asset,
+                year,
+                rep_period,
+                time_block_start,
+                array_agg(variable_index) AS indices,
+                array_agg(duration) AS durations,
+            FROM t_outgoing_nonzero
+            GROUP BY from_asset, year, rep_period, time_block_start
+        ) AS t2
+        ON t1.asset=t2.asset
+            AND t1.year=t2.year
+            AND t1.rep_period=t2.rep_period
+            AND t1.time_block_start=t2.time_block_start
+        ",
+    )
+end
diff --git a/test/test-pipeline.jl b/test/test-pipeline.jl
index 4e09f12c..f331739e 100644
--- a/test/test-pipeline.jl
+++ b/test/test-pipeline.jl
@@ -36,6 +36,7 @@ end
 
     # Create model
     model = create_model(
+        connection,
         graph,
         sets,
         variables,