Graphs. Another way to show SAP ERP data

[aamelin1]

Existing visualisation tools at SAP

How typical SAP report looks like? Like a table with a lot of data or list of entries. Usually, table-like visualisation functionality using to represent data, such as:

• Classic ALV Grid/List - to show transparent table data


• Modern ALV, like IDA, PIVB etc - for tables


• ALV hierarchy, ALV tree - to show hierarchy related data


• FIORI tools to display tables (design studio, lists etc) - for tables too

All these tools are great for displaying table data, but sometimes SAP data is not a table structured, so, it's hard to display this kind of data via these tools.

Non table-like types of SAP data

For example, I think a lot of us make a lot of diagrams with explanation of the business processes with accounts postings as a T-accounting view, like this:

Another example - visualisation of production processes like chain of:

materials->PP orders->semi-finished goods->another PP order->finish good->sales

Like this:

All these data are impossible (or really hard) to represent by table view. Also, existing SAP tools (like FIORI lighthouse app "Display Journal Entries in T-Account View" or SAP Graph) are not so helpful in some cases.So, I think to show these kind of data I should use another approach for visualisation, like, for example Graphs at SAP GUI (SAPLogon) interface.

 

What is Graphs?

Graph is a structure amounting to a set of objects in which some pairs of the objects are in some sense "related". The objects correspond to mathematical abstractions called vertices (also called nodes or points) and each of the related pairs of vertices is called an edge (also called link or line).^[1] Typically, a graph is depicted in diagrammatic form as a set of dots or circles for the vertices, joined by lines or curves for the edges.

Here's a good explanation Wikipedia. Graph (discrete mathematics)

How Graphs can be stored and visualised?

To show Graphs I will use graphviz solution.

Graphviz is open source graph visualization software. Graph visualization is a way of representing structural information as diagrams of abstract graphs and networks.

Just to understand how Graphviz works. I can create a HTML file with simple structure of Nodes and Edges like this:

digraph G {

  a -> b1;

  b1 -> c;

  a -> b2;

  b2 -> c;

}

And open this HTMS file at any browser to get a visualisation:

To store Graph data I'll use a simplest approach - store all nodes and edges at two tables. Another options of storing graph data you may find here 

Technical realisation of graphs at ABAP

How to store graph as an internal table

To store Graph data at SAP I used a two internal tables:

• First one to store Nodes (or vertices). Here should be a unique node ID and some node attributes

Node ID (key)	Name	Other attributes
ID 1	Node one	Color, values, texts etc
ID 2	Node two	...

• Second one - for Edges. Here's a info for relationship between Nodes and additional Edge attributes

Node From (key)	Node to (key)	Edge attributes
ID 1	ID 2	Texts, amounts, color etc

Class to create HTML file of graph

Solution for working with Graphviz at ABAP has been developed by github marcellourbani/abapgraph

I'll use this one, you can install it to your system via ABAPGit or manually (link above).

Simple example how it works:

a) Create a Node structure via se11

b) Create a Edge structure via se11

c) Create a report, select necessary data into two internal tables lt_nodes[] (with p.a structure) and lt_edges[] (with p.b structure), and create a Graph, like this:

*Create Graph

graph = zcl_abap_graph=>create( ).

*Add Nodes

zcl_abap_graph_node_record=>create(...)

*Add Edges (links)

<Node from>-cl->linkto( destination = <Node to> label = ... ).

*Output as HTML file in browser

zcl_abap_graph_utilities=>show_in_browser( graph ).

More details:

TYPES: BEGIN OF t_edge,

           uid TYPE zml_nodes-uid,

           cl  TYPE REF TO zcl_abap_graph_node_record,

         END OF t_edge.



DATA: graph       TYPE REF TO zcl_abap_graph,

      lt_nodes    LIKE SORTED TABLE OF zml_nodes WITH NON-UNIQUE KEY uid,

      lt_edges    LIKE STANDARD TABLE OF zml_edges, 

      gv_html_viewer TYPE REF TO cl_gui_html_viewer,

      cellattrs      TYPE REF TO zcl_abap_graph_attr,

      lt_cl_e        TYPE STANDARD TABLE OF t_edge,

      ls_cl_e        LIKE LINE OF lt_cl_e.

...

*add selection screen

...

*select data into lt_nodes[] and lt_nodes[]

...

*Graph output

TRY.

 graph = zcl_abap_graph=>create( ).

**********************************************************************

*Nodes

 LOOP AT lt_nodes ASSIGNING FIELD-SYMBOL(<nodes>).

        ls_cl_e-uid = <nodes>-uid.

        ls_cl_e-cl = zcl_abap_graph_node_record=>create( 

                     id = |{ <nodes>-uid }| 

                     label =  <nodes>-text_h 

                     graph = graph 

                     escape = abap_false ).

        cellattrs = zcl_abap_graph_attr=>create( abap_true ).

        cellattrs->set( name = 'ALIGN' value = 'left' ).

        CASE  <nodes>-type.

          WHEN  'KALN'. "Materials

            ls_cl_e-cl->addcomponent( 

                     name = 'MatNo:' && |{ <nodes>-matnr ALPHA = OUT }| 

                     value = |{ <nodes>-bwkey }| 

                     nameattributes = cellattrs 

                     valueattributes = cellattrs ).

            ls_cl_e-cl->headerattr->set( name = 'bgcolor' value = pc_KALNC ).

         WHEN 'CORD'. "Co orders

            ls_cl_e-cl->headerattr->set( 

                     name = 'bgcolor' 

                     value = pc_CORD ).

            ls_cl_e-cl->addcomponent( 

                     name =  |{ <nodes>-ktext }| 

                     value = |{ <nodes>-bwkey }| 

                     nameattributes = cellattrs 

                     valueattributes = cellattrs ).

... "Add other Node types

      ENDCASE.

    APPEND ls_cl_e TO lt_cl_e.

  ENDLOOP.

**********************************************************************

*Edges

      LOOP AT lt_edges ASSIGNING FIELD-SYMBOL(<edges>).

        CHECK <edges>-uid_from IS NOT INITIAL.

        CHECK <edges>-uid_to IS NOT INITIAL.

        READ TABLE lt_cl_e ASSIGNING FIELD-SYMBOL(<from>) WITH KEY uid = <edges>-uid_from.

        READ TABLE lt_cl_e ASSIGNING FIELD-SYMBOL(<to>) WITH KEY uid = <edges>-uid_to.

        <to>-cl->linkto( 

                 destination = <from>-cl->id 

                 fontsize = '6' 

                 label = |{ <edges>-tcode  && ':' && <edges>-quant && <edges>-meins }| ).

        UNASSIGN: <to>, <from>.

      ENDLOOP.

**********************************************************************

    zcl_abap_graph_utilities=>show_in_browser( graph ).

  CATCH cx_root INTO ex.

    MESSAGE 'Error' TYPE 'I'.

ENDTRY.

Converting SAP data into a graph

To convert table stored SAP data into Graph, I've used a approach with recursive selection of data. For example, if you need to find all chain of materials (let's call it CostEst -KALNR) and they movements - you may analyse MATDOC table by KALNR field, find all related KALNRs (at field KALNR_CG) and perform again a selection for each found KALNR_CGs as KALNR (and doing it till new KALNR_CG still exist).

Examples how it works at SAP

Case 1. Accounting postings as a T-view

This program selects all FI related data (from ACDOCA etc) and creates a GL accounts as a Nodes and postings as an Edges (based on acdoca-gkont info). As a result - FI documents shows as a T-view accounting entries:

or more complex case with more docs:

Case 2. Production process

This is a program for ML (material ledger) data analysing. Here I selects all ML related data (MLDOC etc) by CostEst numbers and recursive selects all chains of material, CO orders etc. As a result i have here a full list of the production processes.

Legend for the colors below:

• Aquamarine - CO orders


• Lavender - Materials (CostEst)


• Tomato - Material (selection criteria)


• Peach - Outbound Deliveries


• Orange - material revaluation


• Green - Goods receipts

or like this: