2589639 - How does Load work in IQ -- SAP IQ

How loading process works internally

Details on load (pass 1 and pass 2 )

 A load process consists of the following steps:

   a)  read data from file
   b)  convert data from ascii to binary format
   c)  assign rowid to each row of data
   d)  insert data into index
   e)  performing referential integrity checks
   f)   evaluating table and column check constraints

The logic exercised in step (d) varies from index to index.

   -  For  FP(NBit), DATE, DTTM, and TIME indexes, data is inserted directly into the index.
   -  For HG and WD indexes, data is sorted prior to inserting the data into the index. 
   -  Steps (e) and (f) are performed as required.

The above steps are exercised as follows relative to pass 1 and pass 2

  Pass 1
 ======

   steps (a), (b), (c)
   step  (d)  - for  FP(NBit),  DATE, DTTM, and TIME indexes, data inserted into the index
                  - for HG and WD, data put into a sort

  Pass 2
 ======

   step (d)  - for HG and WD indexes, data is retrieved from the sort and inserted into the index
   step (e)  - perform referential integrity checks
   step (f) is performed after Pass 2 is completely finished.

IQ Large memoy (or Virtual memory) is used to store both the data manipulated in step (a) and step (b).  
The amount of virtual memory can be controlled using the "-iqlm" start-up parameter option.  
The sorting done for HG and WD indexes uses IQ temp cache.

Different thread pools are used for Pass 1 and Pass 2. The load process ideally assigns one thread to every IQ index.

SAP IQ15.4
SAP IQ16.x

"Load process" ; "Bulk load" ; , KBA , BC-SYB-IQ , Sybase IQ , How To