Batch Processing

The batch processing functionality enables the execution of large-scale data calculations. Batches consist of sets of commands and algorithms designed to efficiently process not just one or several product items, but typically all items across the network. There are also technical batches that perform auxiliary operations, such as clearing the file cache, training the neural network, etc.

Batch calculations generally require a significant amount of time. Therefore, they are executed either on demand or according to a predefined schedule (for example, price monitoring and forecast calculation batches).

Users are not recommended to manually initiate batch calculations without first reviewing this section. A detailed description of each batch is provided below.

Displays the complete system operation log on the server.

Provides information regarding server memory utilization.

This batch runs a set of tests to evaluate server performance. The system records the execution time of each test and compares it with benchmark values. This batch allows assessment of whether the server running MySales has sufficient performance capacity.

This batch performs a full recalculation of forecasts for all product items. Execution time ranges from 4 to 10 hours, depending on the number of SKUs and stores.

The forecast batch runs automatically overnight according to a schedule agreed upon with the client in order to recalculate forecasts and transfer updated data to the auto-ordering system. If errors occur during scheduled execution, the support team typically launches it manually. Users may also initiate this batch if necessary.

This batch updates competitors’ prices for SKUs included in the price monitoring module. It runs automatically according to the agreed schedule.

This batch updates competitors’ catalogs included in the catalog monitoring module. It runs automatically according to the agreed schedule.

This batch links SKUs in your network with products from competitors’ catalogs obtained during catalog monitoring.

This batch searches for and updates competitors’ prices for SKUs that have registered barcodes in the system. It enables automatic retrieval of prices from competitors that support barcode-based product search and links them to your SKUs using unique barcodes.

This batch initiates neural network training for promo forecasting based on the most recent promo data. It runs automatically once a week according to the agreed schedule. Manual execution must be coordinated with the promo team and the responsible MySales representative.

This batch runs the algorithm for identifying comparable (similar) promos within the network. It is an important batch, as new promo forecasts are based on historical comparable promo data. The system automatically searches for comparable promos when building a new promo forecast, but this batch allows manual execution if needed.

This batch performs a full recalculation of forecasts for all historical promos. Depending on the number of promos, execution may take several days. It is executed only by MySales employees and only in cases where the promo forecasting logic is improved to achieve more accurate results.

This batch recalculates promo forecasts for all promos not older than 14 days, as well as for all future promos. It runs once a week according to the agreed schedule. In case of scheduled execution errors, the support team typically launches it manually. Users may also initiate this batch if necessary.

This batch recalculates promo forecasts for all promos not older than 200 days, as well as for all future promos. It is executed on demand.

This batch deletes all forecasts from the database. It must not be executed by users. It may be performed only by a MySales employee and only in specific situations.

This batch deletes all locally cached historical data. When generating forecasts, MySales caches historical data to avoid reloading it from the database each time. This approach accelerates forecast generation and reduces database load.

If this batch is executed, during the next forecast recalculation MySales will reload all historical data from the database, significantly increasing recalculation time. This batch should be run if changes have been made to historical data.

Similar to the full forecast recalculation batch, this batch calculates forecasts for all SKUs in the network. However, it does not save the calculated forecasts to the database; instead, it collects statistical information available in the “Statistics” section. This allows forecast quality assessment in a shorter time, as saving data to the database is a time-consuming process.

This batch calculates cross-price elasticity for all SKUs in the network. The results are available in the “Price Optimization” section. It runs once a week according to the agreed schedule.

This batch loads all necessary historical data for all items to accelerate the forecasting process. It is logically the reverse of the “Delete ALL File Cache” batch and is recommended after cache deletion.

This batch recalculates group-level sales using SKU-level sales data. It aggregates SKU sales into group sales. It runs once a week according to the agreed schedule. Manual execution is recommended if discrepancies are identified between group-level sales data and aggregated SKU totals.

This batch is available only when using MySales auto-ordering. Its execution automatically recalculates order calendars based on diagrams. While calendars are recalculated automatically when diagrams are configured, users may manually trigger recalculation if necessary.

This batch is available only when using MySales auto-ordering. It automatically recalculates all orders in the system, taking into account the latest forecasts, safety stock, and other factors. It runs daily according to the agreed schedule prior to generating new orders.