# aj Attributes

**URL:** <https://forum.kx.com/t/aj-attributes/10423>\
**Category:** Community Support\
**Tags:** kdb-and-q\
**Created:** [September 18, 2015, 8:24pm UTC](https://forum.kx.com/t/aj-attributes/10423 "2015-09-18T20:24:00Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![r691175002](https://avatars.discourse-cdn.com/v4/letter/r/e56c9b/32.png) [@r691175002](https://forum.kx.com/u/r691175002)\
**Post date:** [September 18, 2015, 8:24pm UTC](https://forum.kx.com/t/aj-attributes/10423/1 "2015-09-18T20:24:00Z")

</div>

I’m hoping to get some clarification on how a splayed table should be arranged. &nbsp;There will be be a maximum of roughly 4 million rows and 600 columns.

This is what I’ve got right now:

c | t f a

---------| ------

date | d s

sym | s id g

open | f

high | f

low | f

close | f

volume | f

vwap | f

However, after reading the aj documentation I noticed that it recommends that the sym column be parted, and that date be sorted “within” sym. &nbsp;As far as I can tell, the s# attribute must be removed from date since it will no longer be ascending for its entire length.

Is my understanding of the documentation correct? &nbsp;How will applying the p# to date affect the speed of other queries? &nbsp;I do not anticipate needing frequent asof-joins and my table is pretty small - which attributes would you recommend I use?

---

<div class="post-metadata">

**Author:** ![LamHinYan](https://avatars.discourse-cdn.com/v4/letter/l/35a633/32.png) [@LamHinYan](https://forum.kx.com/u/LamHinYan)\
**Post date:** [September 18, 2015, 11:35pm UTC](https://forum.kx.com/t/aj-attributes/10423/2 "2015-09-18T23:35:00Z")

</div>

sort by `sym`time, then apply `p# on sym

[http://www.firstderivatives.com/downloads/q\_for\_Gods\_Mar\_2013.pdf](http://www.firstderivatives.com/downloads/q%5C_for%5C_Gods%5C_Mar%5C_2013.pdf)