This is because there would be more than one field common in your two tables so the Qlikview is making synthetic key- to avoid it please make sure there is not more than one field common in both the tables.
In Qlikview script when u don't specify a name to the table, it takes the table name from last exec statement & append to -1,-2 & so on.
Like in your case u r loading summary table after that u have specified another load statement for which there is no table name specified so Qlikview assumes it is part of the same script hence name it to "Summary-1".
SUMMARY: SQL SELECT a.* FROM summary_pcs a; store SUMMARY into E:\qlikview\SUMMARY.qvd (qvd);
you create a QV-table called SUMMARY and store it (exactly: a copy) into a qvd-file. But SUMMARY is still in memory. When loading from the qvd-file
LOAD ........ FROM SUMMARY.QVD (QVD) group by timestamp(EVENT_DATE ......
you read the copy of SUMMARY from the qvd-file into memory, doing some aggregations (sum() etc.). But there exits a table named SUMMARY (the original). So QV calls the "new" one SUMMARY-1 as you didn't specify a new name. And as always QV creates a sync-table with common fieldnames.
Now it is the time to ask you what do you want to achieve at last? Then you can decide wether using a concatenate load or a QV-join to merge SUMMARY and SUMMARY-1 together.Or drop SUMMARY before loading again from qvd. I do not know (yet).
