# Modifying a Single Bin in a Multi-Bin Record

**URL:** <https://discuss.aerospike.com/t/modifying-a-single-bin-in-a-multi-bin-record/12127>\
**Category:** Configuration\
**Created:** [February 28, 2025, 1:05pm UTC](https://discuss.aerospike.com/t/modifying-a-single-bin-in-a-multi-bin-record/12127 "2025-02-28T13:05:47Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![Aadarsh](https://sea1.discourse-cdn.com/flex019/user_avatar/discuss.aerospike.com/aadarsh/32/2560_2.png) [@Aadarsh](https://discuss.aerospike.com/u/Aadarsh)\
**Post date:** [February 28, 2025, 1:05pm UTC](https://discuss.aerospike.com/t/modifying-a-single-bin-in-a-multi-bin-record/12127/1 "2025-02-28T13:05:47Z")

</div>

My access pattern always involves reading the full record, but writes can be divided into logical bins, allowing direct replacement of a bin’s value. I want to understand the performance implications of updating a single bin. I understand that whether updating the entire record or just a single bin, the full record is loaded into memory.

When using the **update** policy for writing back to disk, does Aerospike write only the modified bin or the entire record? Based on this, I want to determine if there is any performance advantage in using separate bins. Otherwise, if the whole record gets written back, I might as well store a key-value map instead in a single bin

---

<div class="post-metadata">

**Author:** ![Tyler\_Ramer](https://sea1.discourse-cdn.com/flex019/user_avatar/discuss.aerospike.com/tyler_ramer/32/2544_2.png) [@Tyler\_Ramer](https://discuss.aerospike.com/u/Tyler_Ramer)\
**Post date:** [March 4, 2025, 1:34am UTC](https://discuss.aerospike.com/t/modifying-a-single-bin-in-a-multi-bin-record/12127/2 "2025-03-04T01:34:39Z")

</div>

Any change to a bin triggers a complete re-write of the record.